大兴安岭典型林分地表死可燃物含水率动态变化及预测模型研究

发布时间：2018-04-07 00:15

本文选题：大兴安岭　切入点：地表死可燃物　出处：《东北林业大学》2016年博士论文

【摘要】：黑龙江省大兴安岭地区是我国面积最大、火灾最严重的林区。如何准确和有效地对该地区森林地表死可燃物含水率进行预测显得尤为重要,同时提高该区地表细小死可燃物含水率预测的准确性对于促进森林防火工作意义重大。在实验室内,对落叶松林、白桦林、落叶松-白桦混交林等3种林型可燃物含水率在不同温、湿度条件下的动态变化进行了测定,定量分析了空气温度和空气相对湿度对可燃物平衡含水率和时滞的影响,选定了适用于本文中可燃物含水率预测模型构建的Nelson和Simard等两种平衡含水率动态方程。以大兴安岭地区西林吉林业局和南瓮河保护区等2个研究区内7种典型林型的地表死可燃物为研究对象,包括山杨-白桦混交林、樟子松林、兴安落叶松林、白桦林、兴安落叶松-白桦混交林、蒙古栎林、沟塘草甸,分析了在春季和秋季防火期内,不同坡向(阴坡和阳坡)、不同坡位(上坡位、中坡位和下坡位)条件下,地表死可燃物含水率和气象要素的动态变化。通过直接估测法(Nelson法和Simard法)和气象要素回归法,构建了各林分地表死可燃物含水率预测模型,根据误差评价了模型预测精度。选择3种典型林型构建了以小时为时间步长的含水率预测模型,分析了在更小时间尺度上建模对预测精度的影响。同时探讨了不同距离气象数据对预测模型精度的影响。主要研究结论如下：1、空气温度与平衡含水率和时滞呈负相关,空气相对湿度与二者呈正相关。Nelson模型和Simard模型对可燃物平衡含水率的拟合效果较好。两种方法适用于本研究中可燃物含水率模型的构建。2、不同条件下可燃物含水率的异质性较强。阴坡落叶松林地表死可燃物含水率为最高,阳坡樟子松林可燃物含水率最低。可燃物含水率与不同气象因子之间呈现显著相关性,并且对不同时长气象因子的响应具有一定的滞后性。3、可燃物含水率预测模型：(1)3种方法中,Simard法模型预测精度最好,气象要素回归法模型预测精度最低。(2)阳坡山杨-白桦混交林模型预测效果最好、沟塘草甸的预测效果最差。(3)秋季防火期模型的预测精度高于春季防火期。(4)在区分不同坡向进行建模时,阳坡林型预测模型预测精度高于阴坡林型。(5)阳坡林分模型的预测精度随坡位的上升逐渐降低,下坡位模型的预测效果最好,阴坡林分与阳坡林分相反,上坡位模型预测精度最高。4、在更小时间尺度上(以小时为步长)构建的可燃物含水率预测模型,精度高于以日为步长的预测模型,更好的满足火行为预报中对地表死可燃物含水率预测精度的要求。5、通过实验样地实测气象数据构建的可燃物含水率预测模型精度要高于使用远距离气象站气象数据构建的模型,能更好的反应可燃物含水率的真实变化情况。综上所述,气象要素、坡向、坡位、林型、以及构建模型的时间尺度等条件均对地表死可燃物含水率的动态变化和模型预测精度产生影响;Nelson法、Simard法和气象要素回归法可在不同精度上对大兴安岭地区典型林分地表死可燃物含水率进行预测,但需根据林型的不同和微地形的变化构建相应的预测模型。
[Abstract]:Greater Khingan Range area of Heilongjiang province is China's largest, the most serious forest fire. How to accurately and effectively on the surface of the forest area dead fuel moisture content forecast is very important, at the same time improve the surface area of fine dead fuel moisture content prediction accuracy of significance to promote the work of forest fire prevention. In the laboratory. The larch forest, birch forest, larch and birch mixed forest 3 types of forest fuel moisture in different temperature, dynamic changes of humidity conditions were determined, quantitative analysis of air temperature and relative humidity of combustible equilibrium moisture content and delay, selected applicable to water combustible paper rate prediction model was constructed by Nelson and Simard and other two kinds of dynamic equations of equilibrium moisture content. In the area of Greater Khingan Range Industry Bureau and Jilin Xilin Nanweng river protection area 2 of 7 kinds of typical The surface dead fuel forest types as the research object, including aspen and birch mixed forest, Pinus sylvestris, larch, birch forest in Xingan, Xingan - Mongolia Larch birch forest, oak forest, meadow pond, analyzed in spring and autumn fire period, different slope (slope and Yang Po), different the slope position (Upper slope, middle slope and lower slope) under the condition of dynamic change of surface dead fuel moisture content and meteorological factors. Through the direct estimation method (Nelson and Simard) and meteorological factors regression method, construct the prediction model of the forest surface dead fuel moisture content, according to the error evaluation model the accuracy of prediction. 3 kinds of typical forest types constructed water hourly time step rate prediction model, analyzes the influence of modeling in smaller time scales on prediction accuracy. We also study the influence of meteorological data on the prediction accuracy of the model in different distance. The main research The conclusions are as follows: 1, the air temperature was negatively correlated with the equilibrium moisture content and time delay, air relative humidity and two were positively correlated.Nelson model and the Simard model is better fitting effect to the combustible equilibrium moisture content. Two kinds of methods suitable for constructing the water fuel ratio model in this study,.2, fuel moisture content in different conditions under the shady slope heterogeneity. Larch surface dead fuel moisture content is the highest, sunny Mongolian pine forest fuel moisture content is lowest. Significant correlation between fuel moisture and different meteorological factors, and has long response of different meteorological factors when lag.3 fuel moisture prediction model: (1) 3 kinds of methods, the model prediction accuracy of Simard method is the best method, meteorological factors regression model prediction accuracy is lowest. (2) the best sunny - Aspen birch forest forecast model, forecasting the meadow pond The worst effect. (3) the prediction accuracy of model was higher than that in spring and autumn fire season fire. (4) in distinguishing different slopes when modeling on forest type prediction model is higher than that of shady slope forest types. (5) prediction model of sunny slope forest gradually decreased along with the rising slope, the best prediction effect a model of the slope, shady slope forest and sunny slope forest on the contrary, the top model prediction accuracy is the highest in.4, smaller time scales (in hours step) water combustible construction rate prediction model has higher accuracy than prediction model to date for the step, meet the fire behavior prediction better on the surface dead fuel moisture content.5 prediction precision, water combustible materials constructed by experimental plots measured meteorological data rate prediction model is more accurate than using remote weather station meteorological data model, the reaction fuel can be better with real water rate Situation. In summary, meteorological factors, slope aspect, slope position, forest type, and construct the model of time scale conditions have an impact on the surface dead fuel moisture content changes and the prediction accuracy of the model; Nelson method in different precision of typical forests in Greater Khingan Range area surface dead fuel moisture prediction rate the Simard method and the meteorological factors regression method, but according to the different forest types and micro topography change prediction model is constructed accordingly.

【学位授予单位】：东北林业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：S762.1

【参考文献】