Research On Prediction Model Of Fire Spread Rate Of Pinus Koraiensis Plantation’s Surface Fuel

Surface fire spread is the main mode of forest fire field expansion,and it is of great significance to explore the development law and mechanism of surface fire spread and construct a prediction model for surface fire spread rate.In this study,60 indoor ignition tests were carried out under simulated field conditions of Pinus koraiensis plantation in Maoershan area of Heilongjiang Province,and 60 indoor ignition tests were carried out under different gradients of fuel moisture content（5%,15%,25%）,load(0.5 kg?m^-2,0.7 kg?m^-2,0.9 kg?m^-2,1.1 kg?m^-2)and slope（0°,10°,20°,30°,40°）,and the spread rate was extracted by thermocouple method.Explore its response relationship under different variable factors.The Rothermel model,Wang Zhengfei model and random forest algorithm were used to predict the fire spread rate,and some parameters of the Rothermel model were corrected,and the prediction accuracy of the four models was compared and analyzed,and the optimal model for predicting the surface fire spread rate of the Pinus koraiensis plantation was evaluated.The main conclusions of this study are as follows:Under various experimental conditions,the moisture content of fuel was significantly negatively correlated with the spread rate of surface fire in Pinus koraiensis plantation,while the load was not significantly correlated with the spread rate,while the slope was significantly positively correlated with the spread rate.There is a significant interaction between slope and water content on the spread rate.In the range of 0°to 20°,the spread rate increases slowly,while in the range of 20°to 40°,the spread rate increases sharply;The promotion effect of slope on the spreading rate will gradually weaken with the increase of water content;Within the range of water content of 5%to 15%,the propagation rate decreases significantly,while within the range of water content of 15%to 25%,the propagation rate decreases slightly.The effect of load on the spread rate has no significant interaction with the moisture content and slope of the fuel.The Rothermel model was used to predict the surface fire spread rate of Pinus koraiensis plantations,with an average absolute error of 0.137 m·min^-1,an average relative error of 38.99%,and an R² of 0.782,but the prediction effect of the spread rate under the conditions of high slope and high moisture content was not satisfactory;After refitting the optimal compression ratio and slope parameters,the Rothermel model had good prediction effect under all experimental conditions,with an average absolute error of 0.045 m·min^-1,an average relative error of 13.37%,and an R² of 0.972;Wang Zhengfei model has a good overall effect in predicting the spread rate under the condition of slope 0°～30°,and the average absolute error is 0.226 m·min^-1,the average relative error is 46.31%,and the R² is 0.468;The random forest model had good prediction effect under all experimental conditions,with an average absolute error of 0.030 m·min^-1,an average relative error of 9.56%,and an R² of 0.991.Among the four prediction models,the Rothermel model after refitting the parameters and the random forest model had good prediction effects and similar prediction accuracy,but the random forest model needed to be further evaluated and verified due to its own characteristics.In general,the modified Rothermel model is more suitable for predicting the surface fire spread rate of Pinus koraiensis plantations at a slope of 0°～40°.