V ol. 40 No. 9Sep. 2020
第40卷 第9期2020年9月中 南 林 业 科 技 大 学 学 报
侧脸图片Journal of Central South University of Forestry & Technology http: //qks.csuft.edu 收稿日期：2019-09-29
基金项目：中央高校基本科研业务费专项资金项目（2015ZCQ-LX-01）；国家自然科学基金项目（U1710123）。
第一作者：马文苑，硕士研究生。 通信作者：冯仲科，教授，博士，博士生导师。E-mail ：******************** 引文格式：马文苑，冯仲科，成竺欣，等.山西省林火驱动因子及分布格局研究[J].中南林业科技大学学报,2020,40(9):57-69.
MA W Y , FENG Z K, CHENG Z X, et al . Study on driving factors and distribution pattern of forest fires in Shanxi province[J]. Journal of Central South University of Forestry & Technology, 2020,40(9):57-69.山西省林火驱动因子及分布格局研究垃圾软件
牙科医师
马文苑1，冯仲科1，成竺欣1，王凤阁2
（1.北京林业大学 精准林业北京市重点实验室，北京 100083；
2. 应急管理部森林防火预警监测信息中心，北京 100054）
摘 要：【目的】研究山西省的林火驱动因子和火险分布格局，可为山西省森林防火工作提供参考。【方法】使用2010—2017年卫星监测热点数据，基于逻辑斯蒂模型和随机森林模型分析气象、地形、植被和人类活动对山西省林火发生的影响，选取山西省主要林火驱动因子，建立林火发生概率模型，并基于最优模型结果绘制山西省森林火险等级区划图，分析山西省森林火险分布格局。【结果】逻辑斯蒂模型选取的山西省主要林火驱动因子有日平均相对湿度、日照时数、日平均气温、日平均风速、海拔、坡度、距道路距离、距居民区距离；随机森林模型选取的山西省主要林火驱动因子有日平均地表气温、日平均气压、日平均相对湿度、日照时数、日平均气温、日平均风速、季度NDVI 和GDP ；逻辑斯蒂模型的预测准确率在84.31%～86.33%之间，随机森林模型的预测准确率在88.98%～94.37%之间。【结论】山西省主要林火驱动因子为气象因子；随机森林模型比逻辑斯蒂模型更适用于山西省林火发生概率的预测；山西省森林火险分布有明显的季节和地域差异，春季的高火险区明显多于其它季节，东部的高火险区多于西部，阳泉市、长治市、晋城市、忻州市东部、晋中市北部、吕梁市东南部和太原市中部是山西省主要高火险区。学钢琴的最佳年龄
英国留学条件关键词：林火驱动因子；逻辑斯蒂回归；随机森林算法；森林火险区划；山西省
中图分类号：S762.2 文献标志码：A 文章编号：1673-923X(2020)09-0057-13
Study on driving factors and distribution pattern of forest fires in Shanxi province
MA Wenyuan 1, FENG Zhongke 1, CHENG Zhuxin 1, WANG Fengge 2
(1. Precision Forestry Key Laboratory of Beijing, Beijing Forestry University, Beijing 100083, China;
2. Forest Fire Prevention & Monitoring Center in Ministry of Emergency Management of China, Beijing 100054, China)
超清美女壁纸
Abstract:【Objective 】The rearch about the driving factors and the distribution pattern of forest fires in Shanxi province can provide information for forest fire prevention.【Method 】Using forest thermal abnormal point data monitored by satellite from 2010 to 2017, logistic model and random forest model, forest fire driving factors in Shanxi province were analyzed and compared bad on four aspects, i.e., climate, topography, vegetation and the human activity. The main forest fire driving factors were identified to establish models of forest fire occurrence probability. Then forest fire-risk zones were mapped bad on the optimal forecast model.【Result 】The main forest fire driving factors identified by logistic model include daily mean relative humidity, sunshine hours, daily mean temperature, daily mean wind speed, elevation, slope, the distance to road and railway and the distance to ttlement. The main forest fire driving factors identified by random forest model include
乘字笔顺daily mean surface temperature, daily mean barometric pressure, daily mean relative humidity, sunshine hours, daily mean temperature, daily mean wind speed, quarterly NDVI and GDP. The forecast accuracy of logistic models is 84.31%-86.33%, and that of random forest models is 88.98%-94.37%.【Conclusion 】Climate factors are the main driving factors of forest fire occurrence in Shanxi province. Random forest model is more suitable than logistic model for forest fire forecast in Shanxi province. There are obvious asonal and regional differences of forest fire risk distribution in Shanxi province. The high fire-risk zones in spring are obviously larger than that in other asons, and that in eastern Shanxi are larger than that in western Shanxi. Yangquan city, Changzhi city, Jincheng city, the eastern part of Xinzhou city, the northern part of Jinzhong city, the southeastern part of Lüliang city and the central part of Taiyuan city are the main high fire-risk zones.
Keywords: forest fire driving factors; logistic regression; random forest algorithm; forest fire risk distribution; Shanxi province
Doi:10.14067/jki.1673-923x.2020.09.007
借鉴的近义词
