Objective  Surface fire is the most common type of forest fire, which directly affects vegetation regeneration and nutrient allocation and circulation of ecosystem. The common indexes reflecting forest fire behavior are fire spreading speed, calorific value per unit area, fire intensity and flame height. The fire behavior simulation based on the actual stand and site conditions can reveal the conditions of forest fire occurrence, effectively judge the possibility of crown fire occurrence, and provide a scientific basis for forest fire prevention and firefighting decision-making.

  Method  Typical forest stands (Robinia pseudoacacia forest, Pinus tabuliformis forest and Platycladus orientalis forest) in Mentougou District of Beijing were selected as the survey objects. 5 sample plots were set for each stand, a total of 15 sample plots. Through field investigation, the data of fuel load (shrub fuel, herb fuel, 1 h fuel load, 10 h fuel load, 100 h fuel load), stand factors (height of the first living branch, height under dead branches, tree height, DBH, canopy density) and site factors (altitude, slope gradient, slope aspect, slope position) were obtained. The BehavePlus6 software was used to simulate the fire behavior indicators of different stand types under varied fuel conditions based on meteorological parameters and fuel parameters, they are the spreading speed of surface fire, the calorific value per unit area, the intensity of fire line and the length of flame. Principal component analysis was conducted with R language, and the potential effects of stand factors, site factors and fuel factors on fire behavior were discussed according to the contribution rate.

  Result  The total fuel loads of Platycladus orientalis forest (POF), Robinia pseudoacacia forest (RPF) and Pinus tabuliformis forest (PTF) were respectively 15.35, 17.59, 15.28 t/ha, in which, the inflammable fuel loads (up-layer leaf, flammable herb, 1 h fuel) were respectively 4.55, 4.41, and 6.18 t/ha, accounting for 29.6%, 25.1% and 40.4% of the total fuel load of the stand, respectively. In the period of fire protection, the average wind speed was 2.2 m/s (corresponding to 7.92 km/h), the simulation result of surface fire rate of spread was PTF > POF > RPF, and the numerical values respectively were 11.5, 11.1, 8.0 m/min. The simulation results of fire heat per unit area were PTF > POF > RPF, and the numerical values were 23 091, 21 155, and 18 413 kJ/m², respectively. The simulation result of fireline intensity was PTF > POF > RPF, and the numerical values were 4 426, 3 882, 2 468 kW/m. The flame height range of POF, RPF and PTF was respectively 0.89−3.40 m, 1.34−2.91 m, 1.78−3.88 m. Under the same conditions, the flame height was PTF > POF > RPF. According to the contribution rates, the principal component analysis results were different by stands. The first and second principal components in POF, RPF and PTF respectively were fuel material composition and stand factors, fuel material composition and fuel moisture content, stand factors and fuel moisture content.

  Conclusion  (1) Flammable combustible material load is a key factor affecting forest fire behavior. (2) Fuel moisture rate plays a decisive role in the value of fire behavior indicators, and the critical value of fuel moisture rate affects the type of forest fire. Combustibles are flammable when dry, and surface fires with fast spread and high intensity are easy to occur in windy weather. (3) The continuity of combustibles is the key factor that determines the development of surface fire into crown fire. The flame height greater than the height under first living or dead branches is very likely to develop from surface fire to crown fire, and it is very difficult to put out. It is recommended to prune and mow regularly to clean up the combustibles under the forest and then reduce the fire risk.