Objective  Canopy fire is a type of high-energy fire which severely damages forest resources. It is difficult to extinguish and threatens the safety of fire fighting personnel. Analyzing the characteristics of canopy combustibles and the occurrence conditions of canopy fires, and simulating potential fire behavior characteristics are of great significance for forest combustibles management and effective prevention and control of canopy fires.

  Method  The study took the Pinus tabuliformis forest in the Beijing Badaling Forest Farm as the research object. Using destructive sampling methods to harvest 18 representative samples of P. tabuliformis, starting from the first living branch height, the canopy of the P. tabuliformis forest was divided from bottom to top with 1 m as a level, those less than 1 m were divided into 1 m level, and investigated the total biomass of combustibles in the canopy according to the diameter of the canopy combustible branches (needles; bough diameter ≥ 0.64 cm; twig diameter < 0.64 cm), which combined the plot area and the average canopy length of the P. tabuliformis forest to calculate the canopy fuel load (CFL) and canopy bulk density (CBD). Based on stand factors, we established a multiple regression model with stand structure parameters (DBH, height of the first living branch, crown length, tree height, crown width); estimated the sample plot based on the canopy fuel load model. Under the conditions of three fine combustibles moisture content (6%, 10%, 14%), the average CBD, combining with the average monthly maximum wind speed and the surface fuel load in the study area, the canopy fire spread rate model of van Wagner and Cruz was used to predict the occurrence of canopy fire in the P. tabuliformis forest, and the Byram model was used to simulate potential fire behavior characteristics (such as the intensity of the fire line and the height of the flame).

  Result  (1) The average canopy fuel load of P. tabuliformis forest was 4.54 t/ha, the CBD was 0.21 kg/m³, and the fuel load distribution was gradually decreasing from bottom to top. The combustibles at the bottom of the canopy (0−1 m) accounted for the largest proportion of the total combustibles in the canopy, which was 54.03%. The boughs distributed at the bottom of the canopy and rapidly decreased layer by layer, and the needles were distributed in a large proportion at each layer. (2)The non-linear model of canopy fuel load based on stand factors had a high degree of fit, in which DBH and height of the first living branch were extremely significantly correlated with CFL (P < 0.01). Under the condition of not destroying the forest, the canopy fuel load of P. tabuliformis forest can be better estimated according to the easy-to-test factor of the forest stand. (3) Under moderate burning conditions, the wind speed was high in spring (March to May), and there was a possibility of continuous canopy fire; under extremely dry and high combustion conditions, the potential fire behavior index of continuous canopy fires from February to May was relatively high. The continuous canopy fire that occurred in April showed the highest potential fire behavior index, with a spreading rate of 46 m/min. The fire line intensity was 8 062 kW/m, and the flame height was 15 m.

  Conclusion  Canopy combustible is an important factor affecting the occurrence of forest fires, and the DBH and the height of first living branch are the main influencing factors of CFL. The measured data of canopy combustibles are directly obtained through destructive sampling, and the constructed canopy fuel load estimation model has high accuracy. Wind speed, CBD and moisture content of fine combustibles are closely related to the occurrence and spread of canopy fires. The forests in the study are prone to high-intensity canopy fires under extreme dry climate conditions. The hidden dangers of canopy fire in the P. tabuliformis forest in spring are greater, high winds and extreme dry climate conditions are prone to high-intensity canopy fires. Through forest tending measures (thinning and pruning), the density of combustibles can be effectively reduced, and the height of live branches can be increased to reduce the probability and harm of canopy fire degree.