大兴安岭雷击火发生的驱动机制及预测模型

doi:10.13287/j.1001-9332.202608.010

应用生态学报 ›› 2026, Vol. 37 ›› Issue (8): 2693-2703.doi: 10.13287/j.1001-9332.202608.010

大兴安岭雷击火发生的驱动机制及预测模型

田晓瑞^1,2*, 陈凤倩³, 宗学政^1,2, 赵凤君^1,2, 李思微^1,2, 邓星越^1,2

¹中国林业科学研究院森林生态环境与自然保护研究所, 北京 100091;
²中国林业科学研究院森林生态环境与自然保护研究所重点实验室, 北京 100091;
³大兴安岭地区加格达奇林业局森林消防一大队, 黑龙江加格达奇 165000

收稿日期:2026-04-23 修回日期:2026-06-16 出版日期:2026-08-18 发布日期:2027-02-18
通讯作者: ^*E-mail: tianxr@caf.ac.cn
作者简介:田晓瑞, 男, 1971年生, 研究员。主要从事森林防火研究。E-mail: tianxr@caf.ac.cn
基金资助:
国家重点研发计划项目(2023YFC3006803, 2023YFD-2202002)资助。

Driving mechanisms and predictive modeling of lightning-caused fires in the Greater Khingan Mountains, China

TIAN Xiaorui^1,2*, CHEN Fengqian³, ZONG Xuezheng^1,2, ZHAO Fengjun^1,2, LI Siwei^1,2, DENG Xingyue^1,2

¹Ecology and Nature Conservation Institute, Chinese Academy of Forestry, Beijing 100091, China;
²Key Laboratory of Ecology and Nature Conservation Institute, Chinese Academy of Forestry, Beijing 100091, China;
³First Forest Firefighting Brigade, Jiagedaqi Forestry Bureau, Daxing’anling, Jiagedaqi 165000, Heilongjiang, China

Received:2026-04-23 Revised:2026-06-16 Online:2026-08-18 Published:2027-02-18

摘要/Abstract

摘要： 为识别大兴安岭雷击火发生的主要驱动因子并构建预测模型,本研究基于2021—2024年闪电、林火及多源环境数据,采用时空匹配筛选候选闪电,针对雷击火发生样本类别不平衡问题,采用集成采样方法构建预测模型样本集,并通过相关性分析、共线性诊断和特征重要性分析筛选建模因子,利用多种统计与机器学习模型开展预测,基于SHAP方法解释最终模型。结果表明:共匹配雷击火闪电103起,闪电-火点平均距离为0.84 km,平均滞留时间为1.4 d,83.5%的火点在3 d内被探测到。预报模型以腐殖质湿度码、干旱码和小时尺度火险指数为核心预报因子,温度和风速为次要因子。多模型比较显示,随机森林模型综合表现最优,其ROC曲线下方的面积(AUC)、敏感性和特异性分别为0.7959、0.7229和0.7752,最优阈值主要集中在0.45～0.55。雷击火发生受火险、气象、地形和闪电特征共同驱动,且具有显著非线性与阈值效应。

关键词: 雷击火, 闪电, 火险天气指数, 可解释机器学习, 大兴安岭

Abstract: To identify the major drivers of lightning-caused fire and to develop predictive models in the Greater Khingan Mountains, we integrated multi-source data from 2021 to 2024, including lightning records, forest fire records, and multi-source environmental data. We identified candidate lightning events using a spatiotemporal mat-ching approach, and addressed the imbalance problem of sample type using the EasyEnsemble method to construct the model dataset. We then selected predictor variables through correlation analysis, collinearity diagnosis, and feature importance analysis, developed multiple statistical and machine learning models, and interpreted the final model by SHAP (SHapley Additive exPlanations). The results showed that a total of 103 lightning events associated with lightning-caused fires were identified. The average distance between lightning strikes and fire ignition points was 0.84 km, with a mean holdover time of 1.4 days. 83.5% of fires were detected within three days after lightning occurrence. The predictive model identified the duff moisture code, drought code, and hourly fire weather index as the primary predictors, with temperature and wind speed as the secondary predictors. The random forest model showed the best overall performance, with an AUC of 0.7959, sensitivity of 0.7229, and specificity of 0.7752. The optimal classification thresholds were mainly concentrated between 0.45 and 0.55. The occurrence of lightning-caused fire was jointly influenced by fire weather, meteorological, topographic, and lightning-related factors, exhibiting significant nonlinear and threshold effects.

Key words: lightning-caused fire, lightning, fire weather index, interpretable machine learning, Greater Khingan Mountains

田晓瑞, 陈凤倩, 宗学政, 赵凤君, 李思微, 邓星越. 大兴安岭雷击火发生的驱动机制及预测模型[J]. 应用生态学报, 2026, 37(8): 2693-2703.

TIAN Xiaorui, CHEN Fengqian, ZONG Xuezheng, ZHAO Fengjun, LI Siwei, DENG Xingyue. Driving mechanisms and predictive modeling of lightning-caused fires in the Greater Khingan Mountains, China[J]. Chinese Journal of Applied Ecology, 2026, 37(8): 2693-2703.

参考文献

[1] Öztürk MG, Bekar ĺ, Tavşanoğlu Ç. Rethinking lightning-induced fires: Spatial variability and implications for management policies. Forest Ecology and Management, 2024, 572: 122262
[2] Janssen TAJ, Jones MW, Finney D, et al. Extratropical forests increasingly at risk due to lightning fires. Nature Geoscience, 2023, 16: 1136-1144
[3] McGuiney E, Shulski M, Wendler G. Alaska lightning climatology and application to wildfire science. Proceedings of the 85th American Meteorological Society Annual Meeting. San Diego, California: 2005
[4] 李威, 舒立福, 王明玉, 等. 1980—2021年大兴安岭山区雷击森林火灾时空分布特征. 林业科学, 2023, 59(10): 22-31
[5] Verjans V, Franzke CLE, Lee SS, et al. Quantifying CO₂ forcing effects on lightning, wildfires, and climate interactions. Science Advances, 2025, 11: eadt5088
[6] Gao C, An R, Wang WQ, et al. Asymmetrical lightning fire season expansion in the boreal forest of Northeast China. Forests, 2021, 12: 1023
[7] Pérez-Invernón FJ, Huntrieser H, Moris JV. Meteorological conditions associated with lightning ignited fires and long-continuing-current lightning in Arizona, New Mexico and Florida. Fire, 2022, 5: 96
[8] Heseltine TD, Abatzoglou JT, Chen Y, et al. Future increases in lightning ignition efficiency and wildfire occurrence expected from drier fuels in boreal forest ecosystems. Environmental Research Letters, 2022, 17: 064008
[9] Moris JV, Álvarez-Álvarez P, Conedera M, et al. A global database on holdover time of lightning-ignited wildfires. Earth System Science Data, 2023, 15: 1151-1163
[10] Hatchett BJ, Nauslar NJ, Brown TJ. Comparing ground-based lightning detection networks near wildfire points-of-origin. Natural Hazards, 2024, 120: 13617-13626
[11] Sharma S, Khanal P. Forest fire prediction: A spatial machine learning and neural network approach. Fire, 2024, 7: 205
[12] Coughlan R, Di Giuseppe F, Vitolo C, et al. Using machine learning to predict fire-ignition occurrences from lightning forecasts. Meteorological Applications, 2021, 28: e1973
[13] Moris JV, Ascoli D, Hunt HGP. A new method to assess the probability of lightning causing wildfires. 37th International Conference on Lightning Protection. Dresden, Germany: 2024: 1072-1077
[14] Shmuel A, Lazebnik T, Glickman O, et al. Global lightning-ignited wildfires prediction and climate change projections based on explainable machine learning models. Scientific Reports, 2025, 15: 7898
[15] 田晓瑞, 舒立福, 赵凤君, 等. 大兴安岭雷击火发生条件分析. 林业科学, 2012, 48(7): 98-103
[16] 王明玉, 袁圣斌, 李威, 等. 大兴安岭密集雷击火过程及影响因子分析. 林业科学, 2022, 58(11): 10-20
[17] 王晓红, 黄颖, 张金玲, 等. 大兴安岭雷击火预测模型研究. 中南林业科技大学学报, 2017(3): 44-48
[18] Gao C, An R, Wang WQ, et al. Lightning-ignited wildfire prediction in northeast China. Agricultural and Forest Meteorology, 2025, 355: 109991
[19] Wang XL, Wotton BM, Cantin AS, et al. Cffdrs: An R package for the Canadian forest fire danger rating system. Ecological Processes, 2017, 6: 5
[20] Van Wagner CE. A method of computing fine fuel moisture content. Ottawa: Canadian Forestry Service, 1977
[21] Tymstra C, Bryce RW, Wotton BM, et al. Prometheus fire growth simulation model. Edmonton: Canadian Forest Service, 2010
[22] Beven K, Kirkby M. A physically based, variable contributing area model of basin hydrology. Hydrological Sciences Bulletin, 1979, 24: 43-69
[23] Kopecký M, Hájek V, číŽková Š, et al. Topographic Wetness Index calculation. Science of the Total Environment, 2021, 755: 142551
[24] Blouin KD, Flannigan MD, Wang XL, et al. Ensemble lightning prediction models for the province of Alberta, Canada. International Journal of Wildland Fire, 2016, 25: 421-432
[25] 汪海潮, 谭玉斌, 石卓, 等. 气溶胶对闪电影响. 大气科学, 2025, 49(5): 1-11
[26] Rorig ML, Ferguson SA. Characteristics of lightning and wildland fire ignition in the Pacific northwest. Journal of Applied Meteorology, 1999, 38: 1565-1575
[27] Kalashnikov DA, Abatzoglou JT, Loikith PC, et al. Lightning-ignited wildfires in the western United States: Ignition precipitation and associated environmental conditions. Geophysical Research Letters, 2023, 50: e2023GL103785
[28] Dorph A, Marshall E, Parkins KA, et al. Modelling ignition probability for human- and lightning-caused wildfires in Victoria, Australia. Natural Hazards and Earth System Sciences, 2022, 22: 3487-3499
[29] Liu XY, Wu JX, Zhou ZH. Exploratory undersampling for class-imbalance learning. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2009, 39: 539-550
[30] Lundberg SM, Lee SI. A unified approach to interpreting model predictions. Proceedings of the 31st International Conference on Neural Information Processing Systems. ACM, 2017: 4768-4777
[31] Moris JV, Conedera M, Nisi LC, et al. Lightning-caused fires in the Alps: Identifying the igniting strokes. Agricultural and Forest Meteorology, 2020, 290: 107990
[32] Wotton BM, Martell DL. A lightning fire occurrence model for Ontario. Canadian Journal of Forest Research, 2005, 35: 1389-1401
[33] 张吉利, 毕武, 王晓红, 等. 雷击火发生的影响因子与预测研究进展. 应用生态学报, 2013, 24(9): 2674-2684

大兴安岭雷击火发生的驱动机制及预测模型

Driving mechanisms and predictive modeling of lightning-caused fires in the Greater Khingan Mountains, China

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	孙海铭, 霍常富, 阴黎明, 何杨辉, 王朋. 大兴安岭多年冻土区森林土壤有机碳分子特征及土壤矿化温度敏感性 [J]. 应用生态学报, 2026, 37(2): 399-408.
[2]	谭洁, 危千骏, 廖朝阳, 邝文俊, 邓慧婷, 余德. 基于XGBoost-SHAP可解释机器学习模型的城市形态与地表温度的关系 [J]. 应用生态学报, 2025, 36(3): 659-670.
[3]	吕晨歌, 贾德彬, 郝玉胜, 尚紫琴, 郭少峰, 谢国英. 内蒙古大兴安岭北部降水同位素特征及水汽来源识别 [J]. 应用生态学报, 2025, 36(3): 859-867.
[4]	沈阳, 李晓英, 蔡慧颖, 许涛, 李景涛, 陈魁. 林火干扰后大兴安岭多年冻土区土壤胞外酶活性变化特征及其影响因素 [J]. 应用生态学报, 2025, 36(2): 497-203.
[5]	纪明宇, 贾德彬, 郝玉胜, 刘婷, 杨丽娜, 李晓艳, 吕晨歌, 尚紫琴. 闪电河流域正蓝旗段水化学和同位素特征及水体转换关系 [J]. 应用生态学报, 2024, 35(11): 3149-3156.
[6]	姜静宜, 孙晓新, 王宪伟, 王淑洁, 马国宝, 陈宁, 杜宇. 大兴安岭多年冻土泥炭地土壤水可溶性有机碳夏秋季节变化特征及其影响因素 [J]. 应用生态学报, 2023, 34(9): 2413-2420.
[7]	董灵波, 梁凯富, 张一帆, 刘兆刚. 基于Landsat 8时间序列数据的翠岗林场森林类型划分 [J]. 应用生态学报, 2022, 33(9): 2339-2346.
[8]	董灵波, 马榕, 田栋元, 王涛, 刘兆刚. 大兴安岭天然林不同演替阶段共优势种种群结构与动态 [J]. 应用生态学报, 2022, 33(8): 2077-2087.
[9]	杨婧雯, 张秋良, 宋文琦, 张旭, 李宗善, 张远东, 王晓春. 大兴安岭兴安落叶松和樟子松径向生长对气候变化的响应差异 [J]. 应用生态学报, 2021, 32(10): 3415-3427.
[10]	及利, 马立新, 程政磊, 祝清超, 张岩, 杨雨春, 杨立学. 大兴安岭北部不同海拔天然林土壤胞外酶化学计量特征及其季节动态 [J]. 应用生态学报, 2020, 31(8): 2491-2499.
[11]	张凌宇, 董灵波, 陈莹, 刘兆刚. 大兴安岭中部天然次生林种子雨动态 [J]. 应用生态学报, 2020, 31(12): 4035-4041.
[12]	梁东哲, 赵雨森, 辛颖. 大兴安岭重度火烧迹地天然次生林土壤温室气体通量及其影响因子 [J]. 应用生态学报, 2019, 30(3): 777-784.
[13]	魏红洋, 董灵波, 刘兆刚. 大兴安岭主要森林类型林分空间结构优化模拟 [J]. 应用生态学报, 2019, 30(11): 3824-3832.
[14]	杨立宾, 隋心, 魏丹, 崔福星, 朱道光, 倪红伟. 大兴安岭棕色针叶林土壤的真菌多样性 [J]. 应用生态学报, 2019, 30(10): 3411-3418.
[15]	张宇婧, 吴志伟, 顾先丽, 付婧婧, 闫赛佳. 火烧强度和火后恢复时间对大兴安岭森林土壤有机碳含量的影响 [J]. 应用生态学报, 2018, 29(8): 2455-2462.