云南松林火险与火行为模型研究

发布时间：2018-06-17 04:20

本文选题：云南松林 + 火险气象预报模型　；参考：《中国林业科学研究院》2017年博士论文

【摘要】：云南松林主要分布于我国西南的云南林区,由于特殊的气候和地理环境,加之火源多发,使得该地区森林火灾高发频发、火情处置困难,急需在该地区开展快速准确的林火预报预测工作。本论文针对云南林区的主要树种云南松,从计算快捷简便的火险气象模型以及预报结果精度高的火行为模拟模型这2个角度入手,收集并整理国内外的主流模型,利用云南松林历史林火资料、历史气象记录、植被数据和地形数据,测算各预报模型在该地区的预报精度,从中筛选适宜云南松林的林火预报预测模型,同时还在测报结果的基础上分析影响预报精度的主要因素,提出模型的改进思路,并在火险气象预报模型的研究结果上研建了云南松林连续化火险气象预报模型以及相应的便于业务部门使用的技术方法。本论文的研究结果可为云南松林的林火预报预测模型选用提供科学依据,研建的连续化火险气象预报模型及相应的技术方法可为该地区的林火预报预测提供技术支撑。具体来看,本论文主要包括以下几个方面的研究内容、方法和结论:(1)森林火灾火险气象预报模型作为当前林火预防的主要技术手段,由于单纯从气象角度进行预报,预报模型计算简单、预报因子获取容易,特别适合在基层业务单位使用。针对当前国内的火险气象预报模型主要都针对东北林区研建的现状,利用云南省曲靖市麒麟区防火办辖区内2013-2015年的实际林火资料,对国内目前主流的火险预报模型进行测报,根据测报结果进行模型评价研究。研究结果表明,“全国森林火险天气等级”的Kappa精度系数值最高,为0.308,其次是“多因子法”的0.233;“全国森林火险天气等级”的EC可接受错误系数位居第3,为0.778,低于“双指标法”、“801法”和“温湿度法”这3个模型的1.000,以及“三指标法”的0.981。因此,“全国森林火险天气等级”是在研究区背景下预报精度最高的火险气象预报模型,适用于云南松林区的森林火险气象预报。(2)针对“全国森林火险天气等级”存在预报结果容易产生阶跃误差、由于直接使用气象台站数据导致预报结果不准确、冗长的计算过程不利于野外防火工作者在巡山护林和扑救指挥时进行快速化的火险计算的问题,提出一种可以克服以上问题的“连续化森林火险气象预报模型”,经过与传统预报结果的比较显示,该改进模型提高了预报精度。同时,基于该改进模型提出了利用GIS系统所提供的小区域地理环境和地图数据分析的功能,进行连续化森林火险气象预报的方法;基于该改进模型提出一种基于嵌入式系统和安卓平台的小区域森林火险等级测报仪,解决以往森林火险预报无法凸显局部重大火险区以及野外防火工作人员以往无法实时探测和计算所在地区火险等级的问题。(3)通过火行为模拟模型获取林火行为特征进行林火预防和扑救是国外林火管理发达国家目前主流采用的林火管理手段。火行为模拟模型需要综合气象、可燃物、地形等多要素,预报结果较之森林火险气象预报准确度提高不少,同时预报结果可同时反映林火发生可能性和发生后的火行为特征,便于为火前预报和火后处置提供决策支持,具有更好的实际参考价值。因此,选用美国和加拿大行业普遍使用的Farsite和Prometheus火行为模拟模型对发生在云南松林的安宁“3·29”森林大火进行模拟,通过对比模拟结果和相关林火资料,定量评价模型的模拟精度。结果表明:在蔓延范围模拟方面,Farsite在Scott可燃物模型下的模拟精度最高,Prometheus最差,但差距不大,Farsite与Prometheus火场范围的差异区主要集中在云南松分布区;在蔓延速度模拟方面,Farsite在2种可燃物模型下的平均ROS模拟输出最接近实际情况,Prometheus则偏离实际情况较远,Farsite与Prometheus的ROS差异区主要集中在云南松分布区;在火线强度模拟方面,Farsite在2种可燃物模型下的平均FLI模拟输出结果类似,Farsite与Prometheus的输出差异较大,差异区主要集中在栎类灌木分布区。
[Abstract]:The Yunnan pine forest is mainly distributed in the Yunnan forest area in the southwest of China. Due to the special climate and geographical environment and the fire source, the forest fire is frequent and difficult to deal with. It is urgent to carry out rapid and accurate prediction of forest fire prediction in this area. This paper is based on the fast calculation of the main tree tree of the Yunnan forest area, Yunnan pine. A simple and convenient fire hazard meteorological model and a high precision simulation model of fire behavior are used to collect and organize the mainstream models at home and abroad, and use the historical forest fire data of Yunnan pine forest, historical meteorological records, vegetation data and terrain data to calculate the prediction accuracy of the prediction model in this area, and select the suitable Yunnan from which Yunnan is suitable. The prediction model of forest fire prediction of the pine forest is also made, and the main factors affecting the prediction accuracy are analyzed on the basis of the results of the report. The improvement ideas of the model are put forward, and the meteorological forecast model of the Yunnan pine forest continuous fire risk and the technical method to be used by the business departments should be built on the result of the study on the weather forecast model of fire insurance. The research results of the paper can provide scientific basis for the selection of forest fire prediction model in Yunnan, and the continuous fire weather forecast model and the corresponding technical methods can provide technical support for the prediction of forest fire in this area. As the main technical means of forest fire prevention, the weather forecast model of forest fire fire risk is the main technical means. Because of the simple weather forecast, the prediction model is simple, the prediction factor is easy to obtain, and it is especially suitable for use in the grass-roots units. Based on the actual forest fire data of 2013-2015 years in the Qilin District fire control area of Qujing, Yunnan, the current mainstream fire risk forecasting model was reported and the model evaluation was carried out according to the results of the report. The results showed that the Kappa precision system of "National Forest Fire Weather Grade" was the highest, followed by "multiple factors". "Law" 0.233; "National Forest Fire Weather Grade" EC acceptable error coefficient is third, 0.778, lower than the "double index method", "801 method" and "temperature and humidity" of the 3 models of 1, and the "three index method" of the 0.981. therefore, "national forest fire danger sky grade" is the highest prediction accuracy in the research area. The weather forecast model of fire insurance is suitable for the weather forecast of forest fire risk in Yunnan pine forest area. (2) the step error is easily produced for the prediction results of "National Forest Fire Weather Grade", because the direct use of meteorological station data leads to the inaccuracy of the forecast results, and the lengthy calculation process is not conducive to the field fire workers in the mountain patrol and the forest protection. The problem of fast fire risk calculation when fighting and commanding is put forward, and a "continuous forest fire weather forecast model" which can overcome the above problems is proposed. After comparing with the traditional forecast results, the improved model improves the prediction accuracy. At the same time, based on the improved model, the small area geography provided by the GIS system is put forward. The function of the environment and the map data analysis to carry out the method of continuous forest fire weather forecast. Based on the improved model, a residential area forest fire risk rating system based on embedded system and Android platform is put forward to solve the previous forest fire risk forecast can not highlight the local major fire danger zone and the field fire workers in the past. The method of real-time detection and calculation of fire risk grade in the region. (3) the forest fire prevention and rescue by fire behavior simulation model is the mainstream method of forest fire management in the developed countries of foreign forest fire management. The simulation model of fire behavior needs many factors such as comprehensive weather, combustibles, terrain and so on. The accuracy of forest fire weather forecast has been improved a lot, and the forecast results can simultaneously reflect the possibility of forest fire and the characteristics of fire behavior after the occurrence of fire. It is easy to provide decision support for pre fire prediction and post fire disposal. Therefore, the Farsite and Prometheus fire, which is widely used in the United States and Canada, are selected. The simulation model was used to simulate the "3. 29" forest fire in Yunnan pine forest. By comparing the simulation results and related forest fire data, the simulation accuracy of the model was evaluated quantitatively. The results showed that the simulation accuracy of Farsite in the Scott combustible model was the highest and the Prometheus was the worst, but the gap was not large, Farsit The difference zone between E and Prometheus fire field is mainly concentrated in the Yunnan pine distribution area; in the spread speed simulation, the average ROS analog output of Farsite under 2 combustible models is most close to the actual situation, Prometheus deviates from the actual situation, and the ROS difference zone between Farsite and Prometheus is mainly concentrated in the Yunnan pine distribution area; it is strong in the fire line. In degree simulation, the average FLI simulation output of Farsite under 2 combustible models is similar. The difference between the output of Farsite and Prometheus is large, and the difference area is mainly concentrated in the distribution area of oak shrubs.
【学位授予单位】：中国林业科学研究院
【学位级别】：博士
【学位授予年份】：2017
【分类号】：S762

【相似文献】