广义Rothermel模型预测平地无风条件下红松-蒙古栎林地表混合可燃物的火行为

Fire behavior of ground surface fuels in Pinus koraiensis and Quercus mongolica mixed forest under no wind and zero slope condition:A prediction with extended Rothermel model

以帽儿山地区红松-蒙古栎林下地表混合可燃物为材料,进行不同含水率、载量和混合比例的室内点烧试验,观测得到蔓延速率、驻留时间、反应强度、火线强度和火焰长度,并与采用表面积加权法和载量加权法的广义Rothermel模型预测值进行比较.结果表明:广义Ro-thermel模型对红松-蒙古栎林下地表混合可燃物的林火蔓延速率、反应强度的预测平均绝对误差分别为0.04m·min-1、77kW·m-2,预测平均相对误差分别为16%、22%;对驻留时间、火线强度和火焰长度的预测偏低,预测平均绝对误差分别为15.5s、17.3kW·m-1和9.7cm,预测平均相对误差分别为55.5%、48.7%和24%.这些误差可以看成是用该模型预测同类可燃物相应火行为的误差下限.两种加权算法对模型预测精度影响差异不大,当红松可燃物所占比重较小时,表面积加权法得到的蔓延速率和反应强度预测值精度较高,载量加权法得到的火线强度和火焰长度预测值精度较高;当红松可燃物所占比重较大时,结果则相反.

A laboratory burning experiment was conducted to measure the fire spread speed, residual time, reaction intensity, fireline intensity, and flame length of the ground surface fuels collected from a Korean pine （Pinus koraiensis ） and Mongolian oak （ Quercus mongolica ） mixed stand in Maoer Mountains of Northeast China under the conditions of no wind, zero slope, and different moisture content, load, and mixture ratio of the fuels. The results measured were compared with those predicted by the extended Rothermel model to test the performance of the model, especially for the effects of two different weighting methods on the fire behavior modeling of the mixed fuels. With the prediction of the model, the mean absolute errors of the fire spread speed and reaction intensity of the fuels were 0.04 m · min-2 and 77 kW· m-2, their mean relative errors were 16% and 22% , while the mean absolute errors of residual time, fireline intensity and flame length were 15.5 s, 17.3 kW · m-1 , and 9.7 cm, and their mean relative errors were 55.5% , 48.7% , and 24%, respectively, indicating that the predicted values of residual time, fireline intensity, and flame length were lower than the observed ones. These errors could be regarded as the lower limits for the application of the extended Rothermel model in predicting the fire behavior of similar fuel types, and provide valuable information for using the model to predict the fire behavior under the similar field conditions. As a whole, the two different weighting methods did not show significant difference in predicting the fire behavior of the mixed fuels by extended Rothermel model. When the proportion of Korean pine fuels was lower, the predicted values of spread speed and reaction intensity obtained by surface area weighting method and those of fireline intensity and flame length obtained by load weighting method were higher; when the proportion of Korean pine needles was higher, the contrary results were obtained.