基于Logistic模型的火后树木死亡率研究进展

Research Progress of Tree Mortality after Fire Based on Logistic Model

Logistic回归模型分析是目前应用最广泛、最有效的预测树木火灾后死亡率的方法。本文通过综述国内外利用logistic回归模型对火灾后的死亡率进行了预测的文献,综合分析模型的建立、评价、使用范围和局限性等,其中至关重要的就是模型中使用的解释变量,树冠损伤和树干损伤是火灾后树木死亡率Logistic回归模型中使用最广泛的解释变量,火灾强度、火灾严重程度(即可燃物消耗程度)、可燃物和火灾类型、燃烧季节、昆虫和病原体、树木活力等都是建立火灾后树木死亡率模型需要考虑的解释变量。综合现有文献,展望未来建立火灾后树木的死亡率预测模型,应该对模型变量进行统一的定义,加入模型验证,并加入与树木死亡率相关的生理反应过程等来增加模型的准确性,使用Logistic对火灾后树木死亡率进行更全面更准确的模拟,以期对火灾后的管理、规划及生态治理提供有价值的参考。

Logistic regression model analysis is currently the most widely used and effective method to predict the mortality of trees after fire.This article summarizes the domestic and foreign literature that uses logistic regression model to predict the mortality rate after fire,and comprehensively analyzes the establishment,evaluation,scope of use and limitations of the model,and the most important thing is the explanatory variables used in the model,canopy damage and trunk damage are the most widely used explanatory variables in the logistic regression model of tree mortality after fire;fire intensity,fire severity(that is,the degree of fuel consumption),combustibles and fire types,burning season,insects and pathogens,tree vitality,etc.are all explanatory variables that need to be considered in establishing a tree mortality model after fire.Integrating the existing literature,looking forward to the establishment of a prediction model for the mortality of trees after fire,the model variables should be defined uniformly,model verification should be added,and the physiological response process related to tree mortality should be added to increase the accuracy of the model.Logistic is used to simulate the mortality of trees after fire more comprehensively and accurately,in order to provide valuable reference for management,planning and ecological governance after fire.