云南松烧除火蔓延速率预测模型研究

Model for predicting the forest fire spreading speed based on the prescriptive combustion of Pinus yunnanensis

基于自制的室内燃烧床模拟野外微缩山地实况进行烧除试验,利用热电偶、秒表、风速仪采集温度、风速、蔓延时间,采用统计软件SPSS用实测的157组林下可燃物含水率x1、载量x_2、可燃物层厚度x_3、坡度x_4、坡向x_5、风速x_6和火蔓延速率y构建云南松烧除火蔓延速率预测的多元线性回归模型。结果表明,该模型拟合度高,可适用于测算云南松烧除火蔓延速率,6个因子的影响程度由大到小为x_6、x_1、x_5、x_4、x_3、x_2,火蔓延速率与风速、可燃物含水率、坡向和坡度相关性高,影响显著。室内模拟烧除试验价格低廉,数据容易采集、操作简便,预测模型简单易懂,能够保证预测数据的时效性和基层烧除实施人员应用。

The paper is concerned about a model for predicting the forest fire spreading speed based on the prescriptive combustion of Pinus yunnanensis,a kind of pine tree. On the condition that the forest burning source can be manually intervened by the prescribed burning to change the stereo structure and the continuous distribution,in hoping to reduce and even avoid the serious forest fire as an efficient and convenient approach to preventing or promoting such fires. Nevertheless,the burning process can itself serve as a cause for unexpected forest fire and result in heavy losses,particularly in the case of tall Pinus yunnanensis with high content of rosin. Therefore,the present paper mainly takes it as its focus on the fire behavior,the spot burning characteristic features and the combustible material management,in addition to the spot burning length,the impact on the local vegetation and soil erosion. Starting from the research focus as is mentioned above,we have collected some combustible forest materials from the sample forest and simulated the situation in an indoor experimental context. At the same time,we have adopted the air-conditioner and humidifier to adjust the temperature and humidity of the forest,with a fan used to imitate the different directions and strengths of the likely wind. We have also made a simulation for the meteorological and topographical situations of the authentic forests in the adjustable experiment field. What is more,we have recorded the temperature,wind speed and spreading time by recording the burning process,and using the thermo-couple,the stopwatch and the anemograph. Thus the experimental data we have gained indicate the regular distribution in the SPSS analysis and the linear regression analysis. Besides,we have built up a multiple linear regression modal by means of the moisture content x1,the loading capacity x_2,the density of combustible materials x_3,slope x_4,aspect x_5,wind speed x_6 and the spreading speed y from the data collected from the 157 sample groups. The results of our simulated research prove that the modal＇s degree of fitting is pretty high with the deviation being low and suitable for testing the fire spreading speed in the abovementioned pine trees. Moreover,we have found that the influential degrees of the six factors can be as follows： x_6 x_1 x_5 x_4 x_3 x_2. Whereas the fire speeding speed is closely related to the moisture content and the hill slope,the indoor simulation experiments tend to be cheaper and easier to conduct and control with the data being collected easier. Thus,it can be concluded that the prediction models prove easy to understand and achieve the time-effectiveness of data and realize the burning plan by grass-root practitioners. Hence,such models can offer reference for the prescribed burning project.