Fire risk zoning in urban areas using logistic regression method (Case study: Kashan city)

Vakilalroaya, Mohammad Amin; Malmasi, Saeed; Zayeem Dar, Mojgan; Mirza Ebrahime Tehrani, Mahnaz

doi:10.30495/girs.2022.685583

Manuscript ID : GIRS-2107-1930 (R1) Visit : 183 Page: 49 - 67

10.30495/girs.2022.685583

http://dorl.net/dor/20.1001.1.26767082.1401.13.4.3.7

Article Type: Original Research

Fire risk zoning in urban areas using logistic regression method (Case study: Kashan city)

Subject Areas : Spatial data infrastructures and standardisation

Mohammad Amin Vakilalroaya ¹ (PhD Student in Environment - Environmental Planning, Faculty of Marine Science and Technology, Islamic Azad University of North Tehran Branch, Tehran, Iran)
Saeed Malmasi ² (Assistant Professor, Department of Environmental Science, Faculty of Marine Science and Technology, Islamic Azad University of North Tehran Branch, Tehran, Iran)
Mojgan Zayeem Dar ³ (Assistant Professor, Department of Environmental Science, Faculty of Marine Science and Technology, Islamic Azad University of North Tehran Branch, Tehran, Iran)
Mahnaz Mirza Ebrahime Tehrani ⁴ (Assistant Professor, Department of Environmental Science, Faculty of Marine Science and Technology, Islamic Azad University of North Tehran Branch, Tehran, Iran)

Received: 2021-07-17 Accepted : 2021-10-04 Published : 2022-10-23

Keywords: Logistic regression, Validation, Fire risk zoning, spatial relationship analysis, Fuzzy,

Abstract :

Background and Objective Fire risk management is a global issue, and urban safety policies must take it seriously. One of the fields of research for controlling urban fires is to identify critical fire points in the region; Insufficient knowledge of these points will cause the occurrence and spread of fire in various areas and uses, delays in controlling it, and causing financial losses and personal injury as well as environmental pollution. Fire risk zoning with the aim of being used in planning and management in urban fire control has not been considered in the study area of this research and no research has been done in this field in the form of research and study plan. The purpose of this study is to determine and identify the criteria for fire risk zoning in the study area, to create a fire risk map based on the logistic regression method and compliance with the fire reality map, and also to present fire management programs and fire crisis management in Kashan.Materials and Methods The steps and techniques used in this study were performed in six steps. The first step is to identify the effective criteria and indicators. Using library studies, information obtained from authoritative articles, and also through the Delphi method in order to collect the opinions of experts, the Likert scale was used. In the second step, the screening of criteria was done in accordance with the purpose. The effective criteria in this study are vulnerable factors including (population density, industrial units, commercial-warehouse, high-rise buildings, old tissue, and fuel station) as well as the capacity of reducing factors. (Fire station, roads, and hydrant valves). In the next step, the layers were standardized using fuzzy logic. At first, the distance function was performed on the criteria in SELVA IDRISI to determine the distance from each phenomenon. Then, by the fuzzy method, all the criteria determined in the range of zero to 255 were standardized. The type of function used in the fuzzy logic approach is linear, and the selection of the function type and thresholds was based on a review of sources and expert opinion. In order to analyze the spatial relationship between fire incidents that occurred in the city and the role of effective factors in its occurrence, all fire points of the last 10 years in the city from 2010 to 2020 were extracted and turned into a raster map. In the fifth step, a Fire hazard map was prepared using logistic regression. After determining the validity of the logistic regression model using the specified indicators, the fire risk zoning map of Kashan was drawn. In the last step, Chi-Square, ROC, and Pseudo R Square were used to validate the logistic regression model.Results and Discussion Advantages of using the logistic regression model In addition to modeling observations, it is possible to predict the probability of each person belonging to each of the levels of dependent variables and the possibility of directly calculating the odds of variables using the maximum likelihood of maximum model coefficients. Also, compared to other statistical techniques, multivariate such as multiple regression analysis and diagnostic analysis, the dependent variable can have only two variables, one is the probability of an accident and the other is its non-occurrence. In order to analyze the spatial relationship between fire incidents that occurred in the city and the role of effective factors in its occurrence, all fire points of the last 10 years in the city from 2010 to 2020 were extracted and turned into a raster map. The output of the logistic regression model has coefficients between 0 and 1, with a probability higher than 0.5 of a value of one (occurrence of fire) and a probability of less than 0.5 of a value of zero (no occurrence of fire) and thus a boolean map of risk is generated. This logarithmic change causes the predicted probability to be continuous in the range of 0 to 1, and the output of the model to be presented as a spatial prediction map of the probability of destruction. Then, in the logistic regression equation, this layer was introduced as a dependent variable and the effective parameters in fire zoning were introduced as an independent variable. After entering the data into the logistic regression statistical model, model coefficients were extracted using effective parameters in IDRISI software. After determining the validity of the logistic regression model using the specified indicators, a fire risk zoning map of Kashan was drawn. Finally, in terms of fire risk potential, the study area was divided into 5 classes: very low, low risk, medium, very high, and high risk. The area of each of the 5 classes was obtained in hectares and the percentages were 87475.47, 4669.03, 132115, 1116.33, 788.96 hectares, and 90.94, 4.85, 2.19, 1.16, 0.82 percent, respectively.Conclusion The value of 0.95 obtained from ROC indicates a very high correlation between the independent and dependent variables. The value of the qi index is twice equal to 110836.07; Since its value is much higher than the threshold value, then the null hypothesis of all coefficients is also rejected. The value of the PR2 test in this study was 0.47, so the logistic regression model had an acceptable fit.

References:

Ahmai ardakani, M. et al (2014). Zoning of forest fire potential areas using multi-criteria decision making methods. Journal of Geography and Environmental Planning. 60 (4): p. 49-66. (in Persian).

Anton, Howard (1994), Elementary Linear Algebra (7th ed.), John Wiley & Sons, pp. 170–171 , ISBN 978-0-471-58742-2

Abedi Gheshlaghi, H., et al (2020). "GIS-based forest fire risk mapping using the analytical network process and fuzzy logic." Journal of Environmental Planning and Management 63(3): 481-499.

Borna, F., et al (2016). "Habitat potential modeling of Astragalus gossypinus using ecological niche factor analysis and logistic regression (Case study: summer rangelands of Baladeh, Nour)". Journal of RS and GIS for Natural Resources, 7 (4): 45-61 (in Persian).

Chuvieco, E. and R.G. Congalton (1989). Application of remote sensing and geographic information systems to forest fire hazard mapping. Remote sensing of Environment. 29(2): p. 147-159.

Chuvieco, E. and J. Salas (1996).Mapping the spatial distribution of forest fire danger using GIS. International Journal of Geographical Information Science. 10(3): p. 333-345.

Chuvieco, E. et al (2010). Development of a framework for fire risk assessment using remote sensing and geographic information system technologies. Ecological Modelling. 221(1): p. 46-58.

Diefenbach, M.A. N.D. Weinstein, and J. O'reilly (1993). Scales for assessing perceptions of health hazard susceptibility. Health education research. 8(2): p. 181-192.

Del Hoyo, L. V., et al (2011). "Logistic regression models for human-caused wildfire risk estimation: analysing the effect of the spatial accuracy in fire occurrence data." European Journal of Forest Research 130(6): 983-996.

Eastman, J. Ronald (2003).IDRISI Kilimanjaro: guide to GIS and image processing, Worcester, MA: Clark Labs, Clark University, pp. 328

Erfani, M., and Ehsanzade, N (2016). " Recreation suitability zoning in part of the Oman sea coast". Journal of RS and GIS for Natural Resources, 12 (1): 107-123 (in Persian).

Habibi, A. A. Sarafrazi, and S. Izadyar (2014). Delphi technique theoretical framework in qualitative research. The International Journal of Engineering and Science. 3(4): p.8-13.

Hosmer DH, Lemeshow S (1989). Applied logistic regression. Wiley series in probability and mathematical statistics. Wiley, New York, 307 p

Jobson, J.D (2012). Applied multivariate data analysis: volume II: Categorical and Multivariate Methods: Springer Science & Business Media.

Juliá, P. B., et al (2021). "Post-earthquake fire risk assessment of historic urban areas: A scenario-based analysis applied to the Historic City Centre of Leiria, Portugal." International Journal of Disaster Risk Reduction 60: 102287.

Jennings, C.R (2013). Social and economic characteristics as determinants of residential fire risk in urban neighborhoods: A review of the literature. Fire Saf. J. 62, 13–19. [CrossRef]

Kosravi, Y. Jabbari, M (2011). Basics of Geographic Information Systems (GIS) and ARC GIS10 tutorial. Azar Kelk Publications: p. 100-150. (in Persian).

Li, S.Y. Tao, G. and Zhang, L.J (2018), "Fire Risk Assessment of High-rise Buildings Based on Gray-FAHP Mathematical Model", Procedia Engineering, Vol 211 No, pp. 395-402.

Lee, S. and B. Pradhan (2006). Probabilistic landslide hazards and risk mapping on Penang Island, Malaysia. Journal of Earth System Science. 115(6): p. 661-672.

Lewis, J.R (1993). Multipoint scales: Mean and median differences and observed significance levels. International Journal of Human Computer Interaction. 5(4): p. 383-392.

Langlois, André (1987). "Clark, WAV et Hosking, PL (1986) Statistical Methods for Geographers. New York, John Wiley and Sons. Cahiers de géographie du Québec 31.82: 91-92.

Martı´nez J, Martı´nez J, Martı´n P (2004). El factor humano en los incendios forestales: Ana´lisis de factores socio-econo´micos relacionados con la incidencia de incendios forestales en Espan˜a. In: Chuvieco E, Martı´n P (eds) Nuevas tecnologı´as para la estimacio´n del riesgo de incendios forestales. CSIC, Instituto de Economı´a y Geografı´a, Madrid, pp 101–142

Martı´nez J, Vega-Garcı´a C, Chuvieco E (2009). Human-caused wildfire risk rating for prevention planning in Spain. J Environ Manag 90:1241–1252

Motavalli, S. Esmaili, R. Hosseinzadeh, M.M (2009), The Signification of Sensitive Regions in the Vaz Catchment by Logistic Regression, Journal of Physiography, Volume 2, Number 5, Autumn, PP. 73-83.

Mrówczyńska, M., et al (2021). "Scenarios as a tool supporting decisions in urban energy policy: The analysis using fuzzy logic, multi-criteria analysis and GIS tools." Renewable and Sustainable Energy Reviews 137: 110598

Papari fard, S. and S Kiani, (2018), Investigation of the vulnerability of the old texture of Bushehr (traditional market area) against fire, civil engineering, Architecture and modern and city administration, Tehran. https://civilica.com/doc/821578 (in Persian)

Sowmya, S. and R. Somashekar (2010). Application of remote sensing and geographical information system in mapping forest fire risk zone at Bhadra wildlife sanctuary, India. Journal of Environmental Biology. 31(6): p. 969.

Schneider, L.C. and R.G. Pontius Jr (2001). Modeling land-use change in the Ipswich watershed, Massachusetts, USA. Agriculture, Ecosystems & Environment. 85(1-3): p. 83-94.

Taridala, S. et al (2017). Expert System Development for Urban Fire Hazard Assessment. Study Case: Kendari City, Indonesia. in IOP Conference Series: Earth and Environmental Science. IOP Publishing.

Vakilalroaya, M. and Zaeimdar, M (2107). The study of fire risk assessment models in urban areas, the third national conference on fire and urban safety. Third National Conference on Fire and Urban Safety, Tehran Municipality Fire and Safety Services. (in Persian).

World Health Organization (2011). Burn Prevention, Success Stories Lessons Learned; World Health Organization: Geneva, Switzerland.

Wei, Y.Y. Zhang, J.Y. and Wang, J (2018). "Research on Building Fire Risk Fast Assessment Method Based on Fuzzy comprehensive evaluation and SVM", Procedia Engineering, Vol 211 No, pp. 1141-1150.

Wang, K., et al (2021). "A POIs based method for determining spatial distribution of urban fire risk." Process Safety and Environmental Protection.

Xin, J. and C. Huang (2013). Fire risk analysis of residential buildings based on scenario clusters and its application in fire risk management. Fire Safety Journal. 62: p. 72-78.

Xu, D. et al (2006). Mapping forest fire risk zones with spatial data and principal component analysis. Science in China Series E: Technological Sciences. 49(1): p. 140-149.

Yagoub, M. and A.M. Jalil (2014). Urban Fire Risk Assessment Using GIS: Case Study on Sharjah, UAE. International Geoinformatics Research and Development Journal, 5 (3): p. 1-8.

Yamagishi, H. and L. Ayalew (2005). "The application of GIS-based logistic regression for landslide susceptibility mapping in the Kakuda-Yahiko Mountains, Central Japan." Geomorphology 65(1-2): 15-31.

Zhang, Y (2013). Analysis on comprehensive risk assessment for urban fire: The case of Haikou City. Procedia Engineering. 52: p. 618-623.

_||_