Study on Pyrolysis Characteristics and Combustibility of Typical Arbor Species Along Different Altitude Gradients in Southwestern Yunnan

Abstract

This study aimed to systematically characterize the pyrolysis characteristics and combustibility of six typical tree species across different altitude gradients in southwestern Yunnan, providing references for fuel management and selection of potential fire-resistant species in this region. Thermogravimetric analysis (heating rate: 20 °C·min⁻¹, air atmosphere) was employed to obtain TG-DTG curves of bark, branches, and leaves. The Coats–Redfern integral method was applied to calculate kinetic parameters, and principal component analysis was conducted for comprehensive combustibility evaluation. The results demonstrated the following: (1) The pyrolysis process of all species underwent the following four distinct stages: moisture evaporation, holocellulose decomposition, lignin decomposition, and ash formation. Among these, holo-cellulose decomposition constituted the primary mass loss stage. Significant differences in pyrolysis characteristics were observed among different plant parts, with leaves and bark exhibiting lower initial pyrolysis temperatures; (2) The activation energy ranged from 56.05 to 86.41 kJ·mol⁻¹ across different components, with branches requiring the highest energy for pyrolysis; (3) Principal component analysis based on multiple indicators yielded the following comprehensive combustibility ranking: Pinus yunnanensis > Betula alnoides > Lithocarpus henryi > Quercus acutissima > Cunninghamia lanceolata > Myrica rubra; and (4) The combustibility assessment results integrating multiple variables (total mass loss rate, stage-specific mass loss, activation energy, and ash content) showed significant differences from the analysis based solely on activation energy, verifying the necessity of a multi-dimensional comprehensive evaluation.

1. Introduction

Wildfires have extensive impacts on forest ecosystems. Low-intensity and periodic prescribed burns can promote the decomposition of organic materials and nutrient cycling, benefiting forest ecosystem health and natural regeneration of forest ecosystems. In contrast, high-intensity or uncontrolled wildfires can incinerate trees, severely damage forest ecosystem stability, and even trigger retrogressive succession [1,2]. A forest fire is a physicochemical reaction process that involves the combination of forest fuels and oxygen at a certain temperature, accompanied by light emission and heat release [3]. Forest fuels serve as the material basis for forest combustion; their combustibility determines the ease of wildfire occurrence and is also a key factor influencing combustion rate and intensity [4]. The pyrolysis characteristics of forest fuels are a crucial component of their combustibility. The initial stage of a wildfire is the thermal decomposition of fuels, which profoundly affects subsequent wildfire behavior [5,6]. Therefore, conducting research on the pyrolysis characteristics of forest fuels is an important foundation for in-depth exploration of wildfire occurrence patterns, fire spread characteristics, and wildfire prevention and suppression [7].

Currently, thermogravimetric analysis (TGA) is the primary method for studying the pyrolysis characteristics and potential combustibility of forest fuels. By continuously measuring the weight change in fuel samples in a controlled temperature-programmed environment, researchers obtain the thermogravimetry (TG) curves and derivative thermogravimetry (DTG) curves of the samples. The pyrolysis process and combustibility of fuels are then analyzed based on these TG-DTG curves [8,9]. Philpot et al. (1970) first proposed using this method to determine the combustibility of plant materials, as it can quickly and accurately characterize the pyrolysis process and thermal stability of solid substances during the pyrolysis stage [10]. This method was introduced to China in the 1990s. Luo et al. (1992) successively used this method to analyze the pyrolysis characteristics of different tree species, providing valuable insights for the selection of fire-resistant tree species [11]. Since then, many scholars have successively applied this method to study the pyrolysis characteristics and pyrolysis kinetics of arbors, shrubs, herbs, and surface litter in different regions. These studies have provided substantial data support for a comprehensive understanding of forest combustibility, in-depth research on the mechanism of fire spread behavior, and exploration of the impacts of wildfires on ecosystems [12,13,14,15,16,17,18,19].

Given the increasing frequency of wildfires in subtropical Asia, understanding the combustibility of local forest fuels in southwestern China is of particular relevance. Southwestern China is rich in forest resources, being the second-largest natural forest area in the country, and is also the region with the highest number of forest fires in China [20]. This region features a variable climate and complex terrain, dominated by high mountain valleys, dry–hot valleys, and plateau mountains. The vegetation exhibits a vertical distribution pattern, mainly consisting of evergreen broad-leaved forests, deciduous broad-leaved forests, and coniferous forests [21]. Therefore, once a wildfire occurs, under the influence of complex terrain and weather conditions, different types of vegetation will exhibit significantly different combustion characteristics and fire behavior patterns [22]. This directly affects the fire spread rate and fireline intensity, posing great challenges to local forest fire suppression efforts. This study focused on the typical arbor species distributed across different altitude gradients in the Pu’er area of southwestern Yunnan, These species were selected for their dominance in local vegetation types and their contrasting fuel structures along altitude gradients. Based on thermogravimetric analysis, this study systematically investigated the pyrolysis behavior of bark, branches, and leaves while conducting a comprehensive assessment of their combustibility. Although these fundamental characteristics do not directly simulate the rapidly changing behavior of real fire scenarios, they constitute the intrinsic determinants of fuel flammability. Consequently, the obtained combustibility ranking of tree species, key pyrolysis parameters, and kinetic data can provide a scientific basis and theoretical reference for evaluating potential fire behavior of fuels, identifying high-risk tree species, and formulating fire prevention strategies in this region.

2. Materials and Methods

2.1. Overview of the Study Area

Pu’er City is located in southwestern Yunnan Province, China (between 22°02′ N and 24°50′ N latitude, and 99°09′ E and 102°19′ E longitude). Its elevation ranges from 317 to 3370 m, with an average of 1302 m. The study area is situated in the southern section of the Hengduan Mountains, where mountain ranges crisscross and vertical climatic differences are significant(Figure 1). It encompasses five climate zones: northern tropical, southern subtropical, central subtropical, northern subtropical, and temperate. The distinct vertical climate zones (ranging from northern tropical to temperate) result in vertical differentiation of vegetation types, providing ideal natural experimental conditions for studying the differences in tree species combustibility across various altitude gradients. The annual average temperature is approximately 19.5 °C, and the annual precipitation ranges from 1200 to 1700 mm (the climate data for this region were sourced from the Pu’er Meteorological Bureau, covering the monitoring period from 2021 to 2023). The region has two distinct seasons: a rainy season (May to October) and a dry season (November to April). During the rainy season, abundant moisture leads to heavy precipitation, with runoff accounting for 70%–80% of the annual total. In contrast, the dry season is characterized by dry air and low precipitation, making it a high-risk period for forest fires. The forest coverage rate in Pu’er exceeds 60%, and even reaches over 70% in some areas, boasting rich biodiversity. The dominant plant families include Fagaceae, Lauraceae, and Magnoliaceae. The main arbor species are as follows: Betula alnoides, Lithocarpus henryi, Quercus acutissima, Cunninghamia lanceolata, Myrica rubra, Pinus yunnanensis, and so on.

2.2. Sample Collection and Preparation

Sample Collection

The sampling layout followed a stratified random sampling strategy. Based on the vertical vegetation zonation of the study area, three elevation gradient zones were clearly delineated: low (500–1500 m), medium (800–2000 m), and high (1500–3000 m). Within each gradient zone, three replicate 20 m × 20 m sample plots were randomly established in pure or dominant stands of each target tree species (Betula alnoides, Lithocarpus henryi, Quercus acutissima, Cunninghamia lanceolata, Myrica rubra, and Pinus yunnanensis) to ensure spatial representativeness and minimize random errors. Within each plot, tree species composition, diameter at breast height (DBH), crown width, and tree height were recorded. Samples of leaves, bark, and branches were collected from the arbor species in each plot. For each species, three replicates per sample type (leaf, bark, branch) were collected, labeled, and weighed, with each sample exceeding 1000 g. All sample collection was completed in December 2024. Detailed information on sample types and their distribution environments is provided in

Table 1. Basic Information of Samples.

Altitudinal Distribution	Sample Name	Mean Tree Height (m)	Mean DBH (cm)	Distribution Environment and Growth Characteristics
Low altitude (ca. 500–1500 m)	Quercus acutissima	11.5 ± 2.1	17.2 ± 3.5	Low mountain and hilly areas, drought-tolerant
	Myrica rubra	8.3 ± 1.7	15.8 ± 2.9	Low mountain and hilly areas, prefers acidic soil
Medium altitude (ca. 800–2000 m)	Lithocarpus henryi	13.8 ± 2.4	21.5 ± 4.2	Subtropical mountainous areas, strong adaptability
	Cunninghamia lanceolata	15.2 ± 2.8	19.3 ± 3.8	Subtropical mountainous areas, prefers warm and humid conditions
High altitude (ca. 1500–3000 m)	Betula alnoides	14.6 ± 2.3	18.7 ± 3.2	Temperate and cold–temperate mountainous areas, prefers humid environment
	Pinus yunnanensis	16.3 ± 2.9	24.8 ± 4.5	Subtropical and temperate mountainous areas, barren soil-tolerant

.

2.3. Sample Preparation

The collected samples were stored in envelopes to maintain dryness, Before the experiment, the samples were oven-dried and dried continuously at 100 °C for 48 h. After moisture content stabilized, they were ground using a WK-1000A oscillating mill (Shanghai Xinnuo Instrument Co., Ltd., Shanghai, China). The crushed samples were then sieved through a 60-mesh sieve (<0.25 mm), and the samples were stored in envelopes for later use, stored in an 80 °C incubator.

2.4. Thermogravimetric Experiment

Thermogravimetric analysis was performed using a TGA55 thermal analyzer (TA Instruments, New Castle, DE, USA). Prior to analysis, the instrument was calibrated for temperature and mass using standard indium and alumina samples. The sample mass used was 10 mg, and the temperature was increased from room temperature to 800 °C at a heating rate of 20 °C·min⁻¹ under an air atmosphere, The flow rate was set at 80 mL·min⁻¹. The instrument was preheated for 20 min before the experiment. All sample measurements were performed in triplicate.

2.5. Data Processing and Analysis

Data processing utilized thermogravimetry (TG) and derivative thermogravimetry (DTG). Origin 2024 was used to plot the TG curves, with temperature as the horizontal axis and sample mass loss rate as the vertical axis, depicting mass change against temperature. The DTG curves, which are the first derivative of the TG curves, were plotted with temperature as the horizontal axis and mass loss rate as the vertical axis, reflecting the rate of mass change during the pyrolysis process. The linear tool in Origin software was used to mark the mass and temperature conditions at each stage of the reaction.

For kinetic data analysis, the Coats–Redfern integral method was adopted. As the TG-DTG curves were obtained at a single heating rate, they meet the application conditions of this method [23]. The relationship between temperature, time, and reaction rate follows the Arrhenius kinetic equation:

$${\displaystyle \frac{d\alpha }{dt}}=kf(\alpha )=A\exp ({\displaystyle \frac{-E}{RT}})f(\alpha )$$

(1)

$$\alpha ={\displaystyle \frac{m_o-m}{m_o-m_{\infty }}}$$

(2)

$$k=A\exp ({\displaystyle \frac{-E}{RT}})$$

(3)

In the above equation: $\mathsf{\alpha }$: conversion rate at a given time; $m_o$: initial mass of the sample (g); m: mass of the sample during pyrolysis at a given time (g); $m_{\infty }$: residual mass of the sample after the completion of thermogravimetric analysis (g); k: Arrhenius rate constant; E: reaction activation energy (kJ·mol⁻¹·K⁻¹); A: pre-exponential factor (min⁻¹); R: universal gas constant (8.314 J·mol⁻¹·K⁻¹); T: pyrolysis kinetic reaction temperature (K).

Integrating Equation (1) gives:

$$G_a={\displaystyle {\int }_0^{\alpha }\frac{d\alpha }{f\left(\alpha \right)}}={\displaystyle {\int }_{T_0}^T\frac{A}{\beta }}\exp \left(-{\displaystyle \frac{E}{RT}}\right)dT$$

(4)

Approximating and taking the logarithm of Equation (4) gives:

$$\ln \left[{\displaystyle \frac{G_{(\alpha )}}{T^2}}\right]=\ln \left({\displaystyle \frac{AR}{\beta E}}\right)-{\displaystyle \frac{E}{RT}}$$

(5)

In Equation (5), β is the heating rate. Based on the basic assumption of the Arrhenius kinetic equation that the reaction rate has an exponential relationship with temperature, the exponential relationship is converted into a linear relationship through logarithmic transformation, facilitating the subsequent determination of activation energy and pre-exponential factor.

Let the equation be:

$$Y=\ln \left[{\displaystyle \frac{G_{\alpha }}{T^2}}\right],a=\ln \left({\displaystyle \frac{AR}{\beta E}}\right),b=-{\displaystyle \frac{E}{R}},X={\displaystyle \frac{1}{T}}$$

Substituting into Equation (5) yields a simplified kinetic equation:

$$Y=a+bX$$

(6)

Plot

$$Y=\ln \left[{\displaystyle \frac{G_{\alpha }}{T^2}}\right]\mathrm{and}X={\displaystyle \frac{1}{T}}$$

in Equation (6) to identify the mechanism function with a relatively high degree of fitting. Based on Equation (6), the frequency factor (A) of the sample can be calculated from the equation of the fitted curve, whereas the activation energy (E) and linear correlation coefficient (R²) can be determined from the slope of the regression line. According to the Arrhenius equation, the magnitude of activation energy reflects the difficulty of the pyrolysis reaction of the substance [24,25], and the thermal stability characteristics of the sample can be judged by the level of activation energy.

Table 2. Common Kinetic Model Functions.

Reaction Mechanism	Symbols	f(α)	G(α)
First-order reaction	F1	1 − α	−ln(1 − α)
Second-order reaction	F2	(1 − α)2	(1 − α) − 1 − 1
Third-order reaction	F3	(1 − α)3	[(1 − α) − 2 − 1]/2
Two-dimensional diffusion	D2	[−ln(1 − α)] − 1	α + (1 − α) ln(1 − α)
Three-dimensional diffusion	D3	[3/2(1 − α)2/3]/[1 − (1 − α)1/3]	[1 − (1 − α)1/3]2
Nucleation and growth	A2	2(1 − α) [−ln(1 − α)]1/2	[−ln(1 − α)]1/2
Nucleation and growth	A3	3(1 − α) [−ln(1 − α)]2/3	[−ln(1 − α)]1/3
Nucleation and growth	A4	4(1 − α) [−ln(1 − α)]3/4	[−ln(1 − α)]1/4
Phase boundary reaction	R1	1	α
Phase boundary reaction	R2	2(1 − α)1/2	1 − (1 − α)1/2
Phase boundary reaction	R3	3(1 − α)2/3	1 − (1 − α)1/3

presents 12 common kinetic model functions [25]. By conducting pyrolysis kinetic analysis on the bark, branch, and leaf samples of all target species, we obtained key parameters for each sample, including the pyrolysis reaction mechanism, activation energy (E), and pre-exponential factor (A). Based on the fitting results, we determined the optimal pyrolysis mechanism by comparing the fitting correlation coefficients of different reaction models. A correlation coefficient closer to 1 indicates better fitting performance. The correlation coefficient (R²) of all samples exceeded 0.99, demonstrating that the Coats–Redfern model can fit our experimental data well and that the selected pyrolysis mechanism model has high reliability. Through kinetic function fitting analysis of the samples during the holocellulose pyrolysis stage, the results showed that the mechanism function with the better fitting effect was the F1 model:

$$G_{(\alpha )}=-\ln (1-\alpha )$$

(7)

The activation energy (E) and pre-exponential factor (A) of different samples varied significantly, indicating corresponding variations in the difficulty of their pyrolysis reactions. The characteristic peaks on the DTG curves (such as the peak temperature and the maximum mass loss rate) intuitively reflect the pyrolysis intensity of fuel components (e.g., hemicellulose, cellulose, and lignin) across different temperature ranges. Meanwhile, kinetic analysis quantifies the energy barriers and molecular collision frequencies of these pyrolysis processes through the activation energy (E) and the pre-exponential factor (A). Together, these two aspects form a complete analytical chain from apparent thermal behavior to intrinsic reaction mechanisms.

2.6. Principal Component Analysis (PCA)

Principal Component Analysis (PCA) is a statistical method that simplifies the analysis process through “dimensionality reduction”—specifically, by selecting a small number of important variables from multiple variables via linear transformation, thereby improving the accuracy of analysis results. When data analysis is conducted with samples grouped, PCA can reflect the original state of the data.

The five original variables (total mass loss rate, holocellulose stage mass loss, lignin stage mass loss, activation energy, and ash content) were standardized using Z-score normalization in SPSS Statistics 26 to eliminate dimensional effects. Principal components with a cumulative variance contribution rate exceeding 85% were retained. The comprehensive scores were calculated using the principal component analysis model, and the combustibility ranking was determined by sorting these comprehensive scores (F) in descending order, where higher scores indicate stronger combustibility [26].

Principal Component Analysis model:

$$F_n={\beta }_1{ZX}_1+{\beta }_2{ZX}_2+\dots +{\beta }_n{ZX}_n$$

(8)

In Equation (8): Fₙ is the score of the nth principal component; β₁, β₂, …, β_n are the eigenvectors corresponding to the eigenvalues of the covariance matrix ∑; ZX₁, ZX₂, …, ZX_n are the z-standardized values of the original variables.

3. Results and Analysis

3.1. Analysis of TG-DTG Curve Characteristics

Figure 2, Figure 3 and Figure 4 illustrate the pyrolysis processes of different plant parts (bark, branch, leaf) of the target arbor species in this study. We observed that the pyrolysis process is mainly divided into four stages, with the main components contributing to mass loss in each stage being moisture, holocellulose, lignin, and ash, respectively.

3.1.1. Analysis of Bark Pyrolysis Characteristics

1.

Moisture Loss Stage

The moisture loss stage occurs from room temperature to approximately 150 °C, with a distinct dehydration peak observed near 75 °C on the DTG curve. The dehydration peak temperatures of barks from different arbor species fall between 68.5 and 83.2 °C. Among them, the bark of Quercus acutissima exhibits the lowest dehydration peak temperature (68.5 °C) on the DTG curve, indicating combustion can initiate at relatively low temperatures. In contrast, the bark of Pinus yunnanensis has the highest dehydration peak temperature (83.2 °C), requiring a higher temperature for combustion initiation.

2.

Holocellulose Decomposition Stage

With further heating, the barks underwent a sharp mass decrease—which constitutes the key stage for the combustion and decomposition of holocellulose in the barks. The holocellulose decomposition peak temperatures range from 255.8 to 293.9 °C. Specifically, the bark of Betula alnoides has the lowest holocellulose decomposition peak temperature (255.8 °C), while the bark of Quercus acutissima has the highest (293.9 °C); the decomposition peak temperatures of the other arbor barks fall between these two values.

In terms of decomposition mass loss: the bark of Myrica rubra shows the smallest mass loss (48.66%), indicating the lowest holocellulose content; the bark of Quercus acutissima has the highest mass loss (63.84%), reflecting the highest proportion of holocellulose content among all samples. The holocellulose mass losses of Pinus yunnanensis, Betula alnoides, Cunninghamia lanceolata, and Lithocarpus henryi barks are 49.69%, 49.93%, 50.15%, and 55.21%, respectively.

The ranking of average holocellulose decomposition rates of barks from different tree species is as follows: Cunninghamia lanceolata bark (3.87%·min⁻¹) < Myrica rubra bark (4.13%·min⁻¹) < Pinus yunnanensis bark (4.18%·min⁻¹) < Betula alnoides bark (4.23%·min⁻¹) < Quercus acutissima bark (4.54%·min⁻¹) < Lithocarpus henryi bark (4.56%·min⁻¹).

As a protective layer of trees, the pyrolysis characteristics of bark are significantly influenced by its lignin, suberin, and ash content. For example, the bark of Quercus acutissima exhibits the highest holocellulose decomposition peak temperature at 293.9 °C, which is closely related to its relatively high suberin content. This type of compound likely enhances the thermal stability of the bark. In contrast, the bark of Betula alnoides shows a decomposition peak at a lower temperature of 255.8 °C and has the highest ash content (2.26%), indicating the presence of more mineral components that catalyze pyrolysis. While this characteristic reduces the ignition temperature, the high ash content inhibits the sustainability of combustion. These findings are consistent with the conclusions of Andreas et al. regarding the fire-resistant properties of bark [27].

3.

Lignin Decomposition Stage

The mass loss peak appearing between 451.5 and 471.1 °C on the DTG curve characterizes the decomposition of lignin. The lignin decomposition peak temperatures of barks from different tree species are in the order: Quercus acutissima bark (451.5 °C) < Cunninghamia lanceolata bark (453.9 °C) < Betula alnoides bark (464.3 °C) < Lithocarpus henryi bark (464.6 °C) < Myrica rubra bark (466.2 °C) < Pinus yunnanensis bark (471.1 °C).

Regarding mass loss: Quercus acutissima bark has the lowest mass loss (33.79%), followed by: Lithocarpus henryi (43.86%), Cunninghamia lanceolata (45.47%), Pinus yunnanensis (45.53%), Myrica rubra (46.07%), and Betula alnoides (46.23%).

In terms of average lignin decomposition rate: Quercus acutissima bark has the lowest rate (4.40%·min⁻¹), while Myrica rubra bark has the highest rate (5.57%·min⁻¹); the average lignin decomposition rates of the other arbor barks fall between these two values.

4.

Ash Formation Stage

The ash content of barks from different arbor species ranges from 1.32% to 2.26%. Among them, the barks of Quercus acutissima and Pinus yunnanensis have the lowest ash content (both 1.32%), while the bark of Betula alnoides has the highest ash content (2.26%), indicating poor combustibility. The ash contents of Cunninghamia lanceolata, Myrica rubra, and Lithocarpus henryi barks are 1.63%, 1.72%, and 2.05%, respectively.

3.1.2. Analysis of Branch Pyrolysis Characteristics

1.

Moisture Loss Stage

The temperature range for moisture loss in branches from different arbor species is 65.5–95.1 °C. Among them, the branch of Pinus yunnanensis has the lowest moisture loss peak temperature (65.5 °C) on the DTG curve, while the branch of Myrica rubra has the highest (95.1 °C). The moisture loss temperatures of Quercus acutissima, Betula alnoides, Lithocarpus henryi, and Cunninghamia lanceolata branches are 79.8 °C, 81.3 °C, 83.2 °C, and 85.7 °C, respectively, falling in the intermediate range.

A comparison of moisture loss between branches and barks reveals significant differences in moisture loss peak temperatures between bark and branches of the same species, indicating variations in water retention capacity between these two plant parts.

2.

Holocellulose Decomposition Stage

The holocellulose decomposition peak temperatures of branches from different arbor species range from 321.4 to 336.2 °C. The branch of Pinus yunnanensis has the lowest holocellulose decomposition temperature (321.4 °C), while the branch of Betula alnoides has the highest (336.2 °C); the branch of Quercus acutissima also has a relatively high decomposition temperature (328.4 °C). Both the bark and branch of Quercus acutissima exhibit high holocellulose decomposition temperatures, suggesting that greater energy requirements for its combustion.

In terms of decomposition mass loss: the mass loss of branches from the six arbor species ranges from 58.12 to 66.50%, and the holocellulose mass loss of the branches is significantly higher than that of the barks. Among them, the branch of Betula alnoides has the highest mass loss (66.50%), while the mass losses of Cunninghamia lanceolata, Quercus acutissima, and Pinus yunnanensis branches are around 58%. The mass losses of Lithocarpus henryi and Myrica rubra branches are 59.49% and 65.74%, respectively.

The average holocellulose decomposition rates of branches from the six arbor species range from 4.67 to 5.45%·min⁻¹, with the following ranking: Pinus yunnanensis branch (4.34%·min⁻¹) < Quercus acutissima branch (4.67%·min⁻¹) < Lithocarpus henryi branch (5.18%·min⁻¹) < Cunninghamia lanceolata branch (5.26%·min⁻¹) < Myrica rubra branch (5.37%·min⁻¹) < Betula alnoides branch (5.45%·min⁻¹). Compared with barks, the holocellulose decomposition rates of the branch and bark of the same tree species also show certain differences.

3.

Lignin Decomposition Stage

The mass loss peak appearing between 449.9 and 470.3 °C on the DTG curve characterizes the decomposition of lignin, consistent with the temperature range of the lignin mass loss peak in barks. The lignin decomposition peak temperatures of branches from different tree species are in the order: Myrica rubra branch (449.9 °C) < Cunninghamia lanceolata branch (451.1 °C) < Betula alnoides branch (462.2 °C) < Quercus acutissima branch (463.5 °C) < Pinus yunnanensis branch (465.7 °C) < Lithocarpus henryi branch (470.3 °C).

In terms of mass loss: the branches of Betula alnoides and Myrica rubra have relatively low mass loss (both below 31%), indicating low lignin content; the branch of Pinus yunnanensis has the highest mass loss (38.24%), reflecting high lignin content; the other tree species fall in the intermediate range.

The average lignin decomposition rates for branches range from 3.53 to 4.75%·min⁻¹. Among them, the branch of Lithocarpus henryi has the lowest decomposition rate (3.53%·min⁻¹), while the branch of Cunninghamia lanceolata has the highest (4.75%·min⁻¹); the lignin decomposition rates of Pinus yunnanensis, Quercus acutissima, Myrica rubra, and Betula alnoides branches are in the intermediate range, at 4.19%·min⁻¹, 4.62%·min⁻¹, 4.04%·min⁻¹, and 3.84%·min⁻¹, respectively.

4.

Ash Formation Stage

The ash content of branches from the six arbor species ranges from 1.19% to 2.85%, which is roughly consistent with the ash content of barks. Among them, the branch of Pinus yunnanensis has a relatively low ash content (1.19%), while the branch of Cunninghamia lanceolata has the highest (2.85%). The ash contents of Betula alnoides, Lithocarpus henryi, Quercus acutissima, and Myrica rubra branches are 1.33%, 1.75%, 2.01%, and 1.53%, respectively.

As a structural fuel, the pyrolysis process of branches is primarily governed by the lignin/cellulose ratio. Pinus yunnanensis branches exhibit the lowest holocellulose decomposition temperature and the highest mass loss during lignin decomposition, which may be attributed to their higher resin content and porous pith structure that facilitate rapid release of volatile matter. In contrast, Betula alnoides branches demonstrate the highest holocellulose decomposition temperature, likely corresponding to their dense xylem structure and higher crystalline cellulose content that require greater energy input to disrupt their crystalline architecture. These differences explain the phenomenon that coniferous branches are more readily involved in combustion during forest fires.

3.1.3. Analysis of Leaf Pyrolysis Characteristics

1.

Moisture Loss Stage

The moisture loss peak temperatures for leaves ranged from 69.2 °C to 83.5 °C. The leaf of Lithocarpus henryi had the lowest peak temperature (69.2 °C), while the leaf of Cunninghamia lanceolata had the highest (83.5 °C). The peak temperatures for Pinus yunnanensis, Quercus acutissima, Myrica rubra, and Betula alnoides leaves were intermediate, at 75.5 °C, 77.6 °C, 81.1 °C, and 81.2 °C, respectively. Significant differences in moisture loss peak temperatures were also observed among the bark, branch, and leaf of the same tree species.

2.

Holocellulose Decomposition Stage

The holocellulose decomposition peak temperatures for leaves ranged from 289.8 °C to 306.1 °C. Compared to branches, the peak temperatures for leaves were significantly lower, indicating that less energy is required for leaf decomposition and that leaves are relatively more flammable. The decomposition peak temperatures for leaves occurred in the following sequence: Cunninghamia lanceolata (289.8 °C) < Pinus yunnanensis (298.1 °C) < Quercus acutissima (298.7 °C) < Myrica rubra (299.5 °C) < Lithocarpus henryi (301.2 °C) < Betula alnoides (306.1 °C). Regarding mass loss, the leaf of Quercus acutissima showed the lowest value (52.38%), indicating low holocellulose content, whereas the leaf of Pinus yunnanensis had the highest (60.12%), suggesting high holocellulose content. The mass losses for Cunninghamia lanceolata, Lithocarpus henryi, Myrica rubra, and Betula alnoides leaves were intermediate at 56.35%, 57.05%, 58.39%, and 59.49%, respectively. Certain differences in holocellulose content were observed among the bark, branches, and leaves of the same species. The average holocellulose decomposition rates for leaves were ranked as follows: Quercus acutissima (4.15%·min⁻¹) < Lithocarpus henryi (4.37%·min⁻¹) < Betula alnoides (4.53%·min⁻¹) < Myrica rubra (4.69%·min⁻¹) < Pinus yunnanensis (4.80%·min⁻¹) < Cunninghamia lanceolata (4.98%·min⁻¹). Overall, the average holocellulose decomposition rate followed the order: bark < leaf < branch.

3.

Lignin Decomposition Stage

The mass loss peak observed between 449.9 °C and 478.8 °C on the DTG curves represents lignin decomposition, consistent with the temperature ranges observed for bark and branches. The lignin decomposition peak temperatures for leaves occurred in this order: Lithocarpus henryi (449.5 °C) < Myrica rubra (452.7 °C) < Cunninghamia lanceolata (463.9 °C) < Betula alnoides (472.3 °C) < Pinus yunnanensis (473.1 °C) < Quercus acutissima (478.8 °C). In terms of mass loss, the leaf of Cunninghamia lanceolata had the lowest value (34.97%), indicating low lignin content, while the leaf of Quercus acutissima had the highest (42.27%), indicating high lignin content. The mass losses for Myrica rubra (36.34%), Pinus yunnanensis (36.91%), Lithocarpus henryi (37.86%), and Betula alnoides (39.86%) were intermediate. The average lignin decomposition rates for the six types of leaves ranged from 2.66%·min⁻¹ to 4.78%·min⁻¹. The leaf of Cunninghamia lanceolata had the lowest rate (2.66%·min⁻¹), and the leaf of Myrica rubra had the highest (4.78%·min⁻¹). The rates for Pinus yunnanensis (3.87%·min⁻¹), Quercus acutissima (4.40%·min⁻¹), Lithocarpus henryi (4.48%·min⁻¹), and Betula alnoides (4.37%·min⁻¹) were intermediate.

4.

Ash Formation Stage

The ash content of leaves from the six tree species ranged from 2.02% to 5.01%. The leaf of Pinus yunnanensis had the lowest ash content (2.02%), and the leaf of Betula alnoides had the highest (5.01%). The ash contents of Myrica rubra, Lithocarpus henryi, Cunninghamia lanceolata, and Quercus acutissima leaves were intermediate, at 2.52%, 3.63%, 4.62%, and 3.16%, respectively.

As key fuels during the initial stages of forest fires, the pyrolysis characteristics of leaves are collectively influenced by specific surface area, wax layers, and secondary metabolites. Pinus yunnanensis needles exhibit distinctive thermal behavior: a prominent holocellulose decomposition peak appears at the relatively low temperature of 298.1 °C, accompanied by a bimodal feature in the DTG curve. This corresponds to the volatilization of abundant terpenoids and epicuticular waxes at lower temperatures, followed by cellulose decomposition at higher temperature ranges. Such thermal characteristics make them highly susceptible to becoming ignition sources under drought conditions. In contrast, the higher ash content (2.52%) and broader decomposition temperature range observed in Myrica rubra leaves reflect the fire-resistant properties typical of evergreen broadleaf species.

3.2. Pyrolysis Kinetics Analysis

The activation energy (E) and pre-exponential factor (A) for different parts of the six tree species were calculated using the Coats–Redfern method, with the results presented in

Table 3. Pyrolysis Parameters of Bark, Branch, and Leaf from Different Arbor Species.

Part	Arbor Species	Fitting Equation	Activation Energy (E) (kJ/mol)	Pre-Exponential Factor (A) (min−1)	Frequency Factor R2
Bark	Betula alnoides	Y = −7876.71X + 0.401	65.49	3.92 × 102	0.99883
	Lithocarpus henryi	Y = −9028.14X + 2.156	75.06	2.6 × 104	0.99907
	Quercus acutissima	Y = −7605.27X − 0.620	63.23	1.36 × 103	0.99803
	Cunninghamia lanceolata	Y = −7694.79X − 0.013	63.97	2.53 × 103	0.99447
	Myrica rubra	Y = 7273.08X − 0.644	60.47	1.27 × 103	0.99076
	Pinus yunnanensis	Y = −9128.42X + 2.444	75.89	3.51 × 104	0.99822
Branch	Betula alnoides	Y = −10,393X + 4.361	86.41	2.71 × 105	0.9933
	Lithocarpus henryi	Y = −9734.19X + 3.483	80.93	1.06 × 105	0.99934
	Quercus acutissima	Y = −8882.19X − 1.967	73.85	2.12 × 104	0.99822
	Cunninghamia lanceolata	Y = −10,131.7X + 4.168	84.24	2.18 × 105	0.99854
	Myrica rubra	Y = −10,336X + 4.417	85.93	2.86 × 105	0.99576
	Pinus yunnanensis	Y = −7845.4X + 0.381	65.23	3.83 × 103	0.99396
Leaf	Betula alnoides	Y = −8557.05X + 1.598	71.14	1.41 × 104	0.99614
	Lithocarpus henryi	Y = −8490.94X + 1.384	70.59	1.13 × 104	0.99892
	Quercus acutissima	Y = −8312.92X + 1.206	69.11	9.26 × 103	0.99783
	Cunninghamia lanceolata	Y = −7682.11X + 0.163	63.87	3.01 × 103	0.99952
	Myrica rubra	Y = −8372.97X + 1.083	69.62	8.24 × 103	0.99851
	Pinus yunnanensis	Y = −6741.59X − 1.388	56.05	5.61 × 103	0.99257

. The R² values all ranged between 0.90 and 0.99, indicating high accuracy of the linear fitting results.

From Table 3, the activation energy of barks during the holocellulose decomposition stage ranges from 60.47 to 75.89 kJ·mol⁻¹. The bark of Myrica rubra has the lowest activation energy, indicating a propensity for combustion at lower temperatures, whereas the bark of Pinus yunnanensis has a relatively high activation energy, suggesting lower flammability under such conditions. The bark of Betula alnoides has the lowest pre-exponential factor, while the bark of Pinus yunnanensis has the highest, implying a higher molecular collision frequency and a more complex pyrolysis reaction requiring greater energy input to sustain. Since activation energy is inversely related to flammability (i.e., lower activation energy corresponds to higher flammability), bark flammability, from lowest to highest, is as follows: Myrica rubra bark < Quercus acutissima bark < Cunninghamia lanceolata bark < Betula alnoides bark < Lithocarpus henryi bark < Pinus yunnanensis bark. Branch activation energy ranges from 65.23 to 86.41 kJ·mol⁻¹. The branch of Pinus yunnanensis has the lowest activation energy, and the branch of Betula alnoides has the highest. The branch of Pinus yunnanensis has the lowest pre-exponential factor, while the branch of Myrica rubra has the highest. Based on the same inverse relationship, branch flammability, from lowest to highest, is ranked as: Pinus yunnanensis branch < Quercus acutissima branch < Lithocarpus henryi branch < Cunninghamia lanceolata branch < Myrica rubra branch < Betula alnoides branch. The activation energy of leaves during the holocellulose decomposition stage ranges from 56.05 to 71.14 kJ·mol⁻¹. The leaf of Pinus yunnanensis has the lowest activation energy, and the leaf of Betula alnoides has the highest. Regarding the pre-exponential factor, the leaf of Cunninghamia lanceolata has the lowest value, and the leaf of Betula alnoides has the highest. Consequently, the flammability of leaves, from lowest to highest, follows the order: Pinus yunnanensis leaf < Cunninghamia lanceolata leaf < Quercus acutissima leaf < Myrica rubra leaf < Lithocarpus henryi leaf < Betula alnoides leaf.

As shown in Table 3, there is a significant positive correlation between the activation energy (E) and the pre-exponential factor (A), indicating the presence of a kinetic compensation effect during the pyrolysis process of the samples. Samples with more complex chemical structures or higher thermal stability tend to compensate through higher effective collision frequencies to maintain considerable reaction rates within specific temperature ranges.

A comprehensive comparison of kinetic parameters across different components (bark, branches, leaves) reveals their connections to forest fire risk: Leaves generally exhibit the lowest activation energy, which aligns with their role as kindling and fire carriers in forest fires. Meanwhile, the higher activation energy of branches suggests their potential contribution as sustained fuels during later stages of fire development, helping to maintain fire intensity and flame height.

3.3. Principal Component Analysis Ordination

Table 4. Combustion Characteristic Parameters of Different Arbor Species.

Part	Samples	X1	X2	X3	X4	X5
Bark	Betula alnoides	97.74	49.93	46.23	65.49	2.26
	Lithocarpus henryi	97.95	55.21	41.18	75.06	2.05
	Quercus acutissima	98.68	63.84	33.79	63.23	1.32
	Cunninghamia lanceolata	98.37	50.15	45.47	63.97	1.63
	Myrica rubra	98.28	49.78	46.07	60.47	1.72
	Pinus yunnanensis	97.98	49.69	45.53	75.89	1.32
Leaf	Betula alnoides	94.99	53.23	39.86	71.14	5.01
	Lithocarpus henryi	96.37	57.05	37.86	70.59	3.63
	Quercus acutissima	96.37	52.38	42.27	69.11	3.16
	Cunninghamia lanceolata	95.38	56.82	34.97	63.87	4.62
	Myrica rubra	97.48	58.39	36.34	69.62	2.52
	Pinus yunnanensis	97.98	60.12	35.98	56.05	2.02
Branch	Betula alnoides	98.47	66.50	30.52	86.41	1.33
	Lithocarpus henryi	98.25	59.49	37.10	80.93	1.75
	Quercus acutissima	97.99	58.39	37.78	73.85	2.01
	Cunninghamia lanceolata	97.15	58.12	35.94	84.24	2.85
	Myrica rubra	98.47	65.74	30.74	85.93	1.53
	Pinus yunnanensis	98.82	58.53	38.24	65.23	1.19

defines the variables as follows: X₁ = total mass loss rate (%), X₂ = mass loss during the holocellulose decomposition stage (%), X₃ = mass loss during the lignin decomposition stage (%), X₄ = activation energy (kJ·mol⁻¹), and X₅ = ash content (%). Among these, X₁–X₃ are positive indicators, whereas X₄ and X₅ are negative indicators.

Principal component analysis (PCA) was employed to comprehensively evaluate the five variables in Table 4 and rank the samples based on their principal component scores. For the six tree species, PCA yielded two principal components with a cumulative contribution rate of 88%. Their expressions are as follows:

Z₁ = 0.41ZX₁ + 0.563ZX₂ − 0.482ZX₃ + 0.359ZX₄ − 0.392ZX₅

(9)

Z₂ = −0.567ZX₁ + 0.287ZX₂ − 0.446ZX₃ + 0.237ZX₄ + 0.584ZX₅

(10)

where Z₁ denotes the first principal component and Z₂ denotes the second principal component.

The comprehensive score (F) was calculated using the formula: F = (49.392 × Z₁ + 35.929 × Z₂)/85.320. The resulting comprehensive flammability scores and ranking for the six tree species are presented in

Table 5. Comprehensive Flammability Scores and Ranking of Different Arbor Species.

Species Name	Bark	Branch	Leaf	Comprehensive Score	Ranking
Pinus yunnanensis	−0.23	0.38	0.39	0.54	1
Betula alnoides	−0.42	1.02	−0.21	0.39	2
Lithocarpus henryi	−0.26	0.43	−0.09	0.08	3
Quercus acutissima	−0.08	0.20	−0.05	0.07	4
Cunninghamia lanceolata	−0.40	0.35	0.01	−0.04	5
Myrica rubra	−0.49	0.39	0.03	−0.07	6

.

Through principal component analysis of the five standardized variables, the first two principal components (Z₁ and Z₂) were extracted. Z₁ shows high positive loadings on mass loss during the holocellulose decomposition stage and mass loss during the lignin decomposition stage, and a negative loading on ash content. It can be interpreted as a ‘comprehensive combustion intensity’ dimension, where a higher score indicates that more combustible mass is released during the pyrolysis process. Z₂ exhibits the highest positive loading on the total mass loss rate and a positive loading on activation energy. It can be interpreted as a ‘combustion efficiency and stability’ dimension, where a higher score suggests that the fuel is more prone to complete combustion and exhibits higher thermal stability. Analysis of the variable loadings reveals that the decomposition amounts of holocellulose and lignin primarily drive Z₁, confirming that structural carbohydrates are key factors influencing combustion intensity. In contrast, the total mass loss rate and activation energy primarily drive Z₂, reflecting the overall combustion efficiency and thermal stability characteristics of the fuel.

From Table 5, the comprehensive flammability ranking (integrating bark, branch, and leaf) of the six tree species is as follows: Pinus yunnanensis > Betula alnoides > Lithocarpus henryi > Quercus acutissima > Cunninghamia lanceolata > Myrica rubra.

Regarding flammability across different plant parts, the comprehensive flammability of branches is generally higher than that of leaves, which in turn is higher than that of bark. Certain interspecific differences were noted; for instance, the leaves of Pinus yunnanensis exhibited stronger flammability compared to leaves of other species, potentially related to the high oil content in the structure of coniferous needles. From the perspective of species distribution across altitudes, the high-altitude species—Pinus yunnanensis and Lithocarpus henryi—showed relatively strong flammability, whereas the low-altitude species—Myrica rubra—exhibited relatively weak flammability across all its parts.

4. Discussion

Thermogravimetric–derivative thermogravimetric (TG-DTG) analysis can comprehensively characterize the combustion performance of different tree parts. In this study, the pyrolysis processes of different parts across species showed similar patterns, primarily divided into four stages: moisture loss, holocellulose pyrolysis, lignin pyrolysis, and ash formation. This is consistent with the findings of Chen et al. and Guo et al. [28]. Among these, the holocellulose decomposition stage was the main mass-loss stage, with mass loss rates generally exceeding 49%, indicating that holocellulose dominates the initial combustion stage and is a primary component of forest fuel combustion [29].

However, significant differences in pyrolysis characteristics were observed among parts: leaves and bark showed a stronger combustion tendency due to their lower initial pyrolysis temperature and higher volatile matter content, which aligns with the results of Zhai and Liu (2008) [24]. The activation energy of branches was generally higher, indicating greater energy requirements for pyrolysis and stronger thermal stability. In contrast, the lower activation energy of leaves may be related to secondary cracking and dehydrogenation reactions, facilitating ignition—a finding consistent with Niu et al. [30].

The observed differences in activation energy in this study primarily stem from the heterogeneity in the biochemical composition of the fuels. For instance, the bark of Myrica rubra exhibits a relatively high activation energy, which may be attributed to its higher lignin and mineral (ash) content. The thermal decomposition of these components requires overcoming higher energy barriers. Conversely, the extremely low activation energy of Pinus yunnanensis needles aligns with the characteristics of their abundant volatile extracts (such as resins), which are prone to decompose and combust at lower temperatures. The significant positive correlation between E and A reveals a typical kinetic compensation effect. This implies that samples with more complex decomposition pathways (higher E) compensate through higher molecular collision frequencies (higher A), thereby maintaining relatively stable apparent reaction rates within the experimental temperature range.

Reported activation energy values in different studies range from 20 to 200 kJ·mol⁻¹, with the difference between maximum and minimum values not exceeding one order of magnitude. In contrast, the pre-exponential factor values vary significantly, differing by up to two orders of magnitude across species and parts. These considerable variations may be attributed to differences in fuel chemical composition, model selection for kinetic analysis, parameter estimation methods, and inherent limitations of the models. Future research should further investigate the causes of these differences [13].

A comparison between the flammability ranking based solely on activation energy and the comprehensive ranking from principal component analysis (PCA) revealed the limitations of a single index (e.g., activation energy) in fully evaluating fuel flammability. This study adopted PCA, integrating multiple variables—including total mass loss rate, stage-specific mass loss, activation energy, and ash content—enabling a more systematic evaluation. The results differed from the ranking based on a single kinetic parameter, further verifying the necessity of a multi-dimensional comprehensive evaluation [30,31,32,33,34].

This study confirms that differences in flammability among species result from the combined effects of thermodynamic parameters and biochemical composition. High contents of volatile extracts and resins (as found in Pinus yunnanensis) significantly reduce the ignition temperature of fuels and enhance the potential for fire spread. In contrast, a high proportion of lignin and mineral components (as seen in Myrica rubra) improves fuel’s fire resistance by enhancing thermal stability and forming an insulating ash layer. Therefore, incorporating fuel chemical composition data into future fire prediction models is expected to enhance forecasting accuracy.

Based on the comprehensive species evaluation, high-altitude species such as Pinus yunnanensis and Betula alnoides possess stronger flammability. Once ignited, fires in these stands could result in high intensity, underscoring the need for enhanced fuel management, such as branch pruning and surface litter clearance. Conversely, all parts of the low-altitude species Myrica rubra exhibited weaker flammability and good fire resistance, making it suitable as a fire-resistant species for stand transformation or firebreak establishment to reduce overall fire risk. Comparisons show that vegetation flammability varies significantly across regions; therefore, fire prevention measures should be selected in a targeted manner, tailored to local conditions [35,36,37,38,39,40,41,42,43].

5. Conclusions

This study presents a systematic investigation into the pyrolysis characteristics and combustibility of six typical arbor species across different altitude gradients in the Pu’er region of southwestern Yunnan. The primary innovation lies in the integrated application of thermogravimetric analysis, kinetic modeling, and principal component analysis to comprehensively evaluate the fuel properties of various plant parts (bark, branch, leaf) under complex altitudinal conditions. The main conclusions are as follows:

(1)

The pyrolysis process of combustible materials is divided into four stages: moisture loss, holocellulose pyrolysis, lignin pyrolysis, and ash formation. The holocellulose decomposition stage accounted for the greatest mass loss, with rates exceeding 49%. Significant differences in pyrolysis characteristics were observed among parts: leaves and bark had lower initial pyrolysis temperatures and higher volatile matter content, presenting a higher combustion risk. The average holocellulose decomposition rate followed the order: bark < leaf < branch. The ash content of different parts ranged from 1.19% to 5.01%.

(2)

The activation energy during the holocellulose decomposition stage ranged from 60.47 to 75.89 kJ·mol⁻¹ for bark, 65.23 to 86.41 kJ·mol⁻¹ for branches, and 56.05 to 71.14 kJ·mol⁻¹ for leaves. This indicates that branches require the highest energy for pyrolysis, followed by bark, with leaves requiring the least.

(3)

The comprehensive combustibility ranking derived from principal component analysis—integrating total mass loss rate, stage-specific mass loss, activation energy, and ash content—was: Pinus yunnanensis > Betula alnoides > Lithocarpus henryi > Quercus acutissima > Cunninghamia lanceolata > Myrica rubra. Beyond providing a combustibility order, this study offers mechanistic insights into the underlying causes, such as the presence of kinetic compensation effects and the antagonistic roles of volatile extracts versus ash content in influencing flammability.

(4)

High-altitude tree species such as Pinus yunnanensis and Betula alnoides exhibit strong flammability, necessitating enhanced fuel management strategies in high-elevation forests—such as branch pruning and litter clearance—to mitigate fire risk. In contrast, the low-altitude species Myrica rubra demonstrates consistently low flammability across all components, making it suitable for recommendation as a fire-resistant species for stand transformation or firebreak establishment in this region.

The findings of this study provide references for identifying fire-prone tree species, implementing differentiated fuel management, and strategically selecting fire-resistant species in southwestern Yunnan, while also offering data support for the parameter localization of regional forest fire behavior models. Subsequent research could integrate field burning experiments to further validate the relationship between laboratory indicators and actual fire behavior.