| 122 | 0 | 37 |
| Downloads | Citas | Reads |
To strengthen fire prevention in mountainous tourist cities, this study proposes a systematic prevention and control method that integrates fire prediction with prevention optimization. Huangshan City, a renowned tourist destination in China characterized by over 50% mountainous terrain and frequent fire risks associated with tourism activities, was selected as a case study. A comprehensive dataset was constructed by integrating fire hotspot data from the Moderate-resolution Imaging Spectroradiometer(MODIS) with multi-source geospatial data, including meteorological, topographic, vegetation, and urban planning information. Two deep learning models—a Convolutional Neural Network(CNN) and a Long Short-Term Memory(LSTM) network—were developed to perform spatiotemporal predictions of fire occurrence probability at a spatial resolution of 1 km × 1 km and on a daily temporal scale. Model performance was evaluated using multiple metrics: accuracy, precision, recall, and the Area Under the Receiver Operating Characteristic Curve(AUC-ROC). The LSTM model significantly outperformed the CNN model, achieving an accuracy of 88%, a precision of 87%, a recall of 90%, and an AUC-ROC of 0.95, demonstrating its superior suitability for fire prediction in the study area. Sensitivity analysis of the driving factors revealed that daily maximum temperature was the key driver of fire occurrence, contributing 35.2%, while wind speed had a relatively minor impact, contributing only 8.7%. This pattern differs significantly from those observed in urban areas. Using Geographic Information System(GIS) spatial analysis, the fire risk sensitivity of Huangshan City was quantified and zoned, revealing a distinct pattern of higher risk in the central-western regions and lower risk in the northern and southern areas. Subsequently, a location-allocation model was employed to optimize the layout of fire stations with the goal of maximizing fire emergency coverage within a 5-minute response time. The optimized layout scheme recommends retaining 5 existing fire stations and adding 23 new stations in high-risk and underserved areas. Verification of the optimized scheme indicates that the overall fire emergency coverage rate for Huangshan City increased from 59.27% to 80.53%. By integrating qualitative and quantitative fire risk analyses, this method effectively enhances the scientific foundation for pre-disaster prevention and improves the efficiency of post-disaster emergency response in mountainous tourist cities.
[1] 高明明，王竹，裘知，等.山地农村火灾分布特征及村落因素的相关性分析：以黔东南为例[J].灾害学，2023,38(3):100-104,147.Gao Mingming,Wang Zhu,Qiu Zhi,et al.Distribution characteristics of mountainous rural fire and its correlation with village factors:take southeastern Guizhou as an example[J].Journal of Catastrophology,2023,38(3):100-104,147.
[2] Yoshiaki O.A primitive study of a fire detection method controlled by artificial neural net[J].Fire Safety Journal,1991,17(6):535-553.
[3] 吴龙标，邓超.火灾探测的人工神经网络方法[J].人类工效学，1997,3(2):39-41,71.Wu Longbiao,Deng Chao.Artificial neural networks in fire detection[J].Chinese Journal of Ergonomics,1997,3(2):39-41,71.
[4] Silva D,Ferreira T M,Rodrigues H.Wildland-urban interface fire exposure of rural settlements:the case of Montesinho Natural Park[J].International Journal of Disaster Risk Reduction,2024,112:104790.
[5] 张金钰，彭道黎，张超珺，等.基于深度学习的内蒙古大兴安岭林区火灾预测建模研究[J].林业科学研究，2024,37(1):31-40.Zang Jingyu,Peng Daoli,Zhang Chaojun,et al.Fire prediction modeling in the Greater Khingan Mountains forest region of Inner Mongolia based on deep learning[J].Forest Research,2024,37(1):31-40.
[6] 姜学鹏，刘飞扬，王启睿，等.基于MCLP-BACOP模型的城市消防站选址研究[J].火灾科学，2022,31(4):252-260.Jiang Xuepeng,Liu Feiyang,Wang Qirui,et al.Study on site selection of urban fire station based on MCLP-BACOP model[J].Fire Safety Science,2022,31(4):252-260.
[7] 邹亮.遥感与GIS在消防规划火灾风险评估中的应用[J].数字通信世界，2011,74(2):67-71.Zou Liang.Application of remote sensing and GIS in fire risk assessment of fire planning[J].Digital Communication World,2011,74(2):67-71.
[8] Shahparvari S,Fadaki M,Chhetri P.Spatial accessibility of fire stations for enhancing operational response in Melbourne[J].Fire Safety Journal,2020,117:103149.
[9] 祝明明，罗静，余文昌，等.城市POI火灾风险评估与消防设施布局优化研究：以武汉市主城区为例[J].地域研究与开发，2018,37(4):86-91.Zhu Mingming,Luo Jing,Yu Wenchang,et al.Urban fire risk evaluation and location optimization of fire station based on the POI:a case study of main urban region in Wuhan[J].Areal Research and Development,2018,37(4):86-91.
[10] 王爱，张强，陆林，等.多源数据支持下城市火灾风险评估及规划响应[J].中国安全科学学报，2021,31(3):148-155.Wang Ai,Zhang Qiang,Lu Lin,et al.Urban fire risk assessment and planning response based on multi data[J].China Safety Science Journal,2021,31(3):148-155.
[11] 晏风，袁沙沙，徐匆匆，等.基于GIS与网络分析协同的城市火灾风险概率评估[J].安全与环境学报，2022,22(3):1147-1153.Yan Feng,Yuan Shasha,Xu Congcong,et al.Study on the probability assessment of urban fire risk based on GIS coupled with network analysis[J].Journal of Safety and Environment,2022,22(3):1147-1153.
[12] 吴运法，王岩青，陈萨如拉，等.集聚型传统村落火灾蔓延模型构建与路径分析：以安徽宏村为例[J].清华大学学报(自然科学版),2025,65(4):644-654.Wu Yunfa,Wang Yanqing,Chen Sarula,et al.Construction and path analysis of fire spread models in traditional village clusters:a case study of Hongcun in Anhui Province[J].Journal of Tsinghua University (Science and Technology),2025,65(4):644-654.
[13] 胡久祎，秦广志，白建国，等.徽州传统古村落消防保护现状及建议[J].建筑安全，2023,38(11):35-37,42.Hu Jiuyi,Qin Guangzhi,Bai Jianguo,et al.Current status and suggestions on fire protection of Huizhou traditional ancient villages[J].Construction Safety,2023,38(11):35-37,42.
[14] 建标152—2017 城市消防站建设标准[S].Construction Standard 152—2017 Standard for construction of urban fire stations[S].
[15] 闫金花.山地城市商业建筑火灾控制模糊系统分析[J].消防科学与技术，2015,34(9):1254-1257.Yan Jinhua.Analysis of fire control in business buildings in mountain cities by fuzzy system[J].Fire Science and Technology,2015,34(9):1254-1257.
[16] 曹丽丽，刘向宇，陈响，等.吉林省延边地区森林火灾发生概率预测模型[J].东北林业大学学报，2024,52(3):90-96.Cao Lili,Liu Xiangyu,Chen Xiang,et al.Prediction model of forest fire occurrence probability in Yanbian Area,Jilin Province[J].Journal of Northeast Forestry University,2024,52(3):90-96.
[17] 郗婕，傅微.基于机器学习的流域尺度森林火灾灾害风险预测[J].自然灾害学报，2024,33(1):89-98.Xi Jie,Fu Wei.Watershed-scale forest fire risk prediction based on machine learning[J].Journal of Natural Disasters,2024,33(1):89-98.
[18] 徐智邦，周亮.基于POI数据的巨型城市消防站空间优化：以北京市五环内区域为例[J].地理科学进展，2018,37(4):535-546.Xu Zhibang,Zhou Liang.Spatial optimization of mega-city fire station distribution based on Point of Interest data:a case study within the 5th Ring Road in Beijing[J].Progress in Geography,2018,37(4):535-546.
[19] 朱龙祥，王自法，张昕，等.基于GIS和多种机器学习算法的广东省森林火灾预测模型[J].林业工程学报，2024,9(3):159-167.Zhu Longxiang,Wang Zifa,Zhang Xin,et al.Forest fire prediction models for Guangdong Province based on GlS and multiple machine learning algorithms[J].Journal of Forestry Engineering,2024,9(3):159-167.
[20] Wang Kai,Yuan Yingfeng,Cheng Mengmeng,et al.A POIs based method for determining spatial distribution of urban fire risk[J].Process Safety and Environmental Protection,2021,154:447-457.
[21] Sharama L K,Gupta R,Fatima N.Assessing the predictive efficacy of six machine learning algorithms for the susceptibility of Indian forests to fire[J].International Journal of Wild land Fire,2022,31(8):735-758.
[22] 胡嫚莉，雍新琴，李鑫，等.城乡融合发展导向下农村居民点空间布局优化研究：以徐州市铜山区为例[J].地理与地理信息科学，2022,38(5):104-110.Hu Manli,Yong Xinqin,Li Xin,et al.Spatial optimization of rural residential area from the perspective of urban-rural integration development:a case study of Tongshan district[J].Geography and Geo-Information Science,2022,38(5):104-110.
Basic Information:
DOI：10.13637/j.issn.1009-6094.2025.1709
China Classification Code:TP18;TU998.1
Citation Information:
[1]Wu Yunfa,Cheng Yuqi,Chen Sarula ,et al.Fire prediction and prevention optimization for mountainous tourist cities based on deep learning[J].Journal of Safety and Environment,2026,26(09):3393-3403.DOI:10.13637/j.issn.1009-6094.2025.1709.
Fund Information:
国家自然科学基金面上项目(52478001); 安徽省高校自然科学研究项目-重点项目(2024AH050240); 国家消防救援局科技计划项目(2023XFLR45)
2026-07-14
2026-07-14
2026-07-14