Comparative Analysis of Potential Fire Behavior Among Three Typical Tree Species Fuel Loads in Central Yunnan Region

Abstract

Potential fire behavior varied significantly among tree species, directly influencing forest fire intensity and spread. To quantify these differences and evaluate species-specific fuel traits for fire management applications, this study conducted field surveys and sample collection in the Jin Dian Yuanbaoshan Forest Area, Kunming, Yunnan Province. Using a combustion bed experiment, we simulated the burning behavior of Acacia dealbata, Alnus nepalensis, and Pinus armandii under windless conditions, recording ignition time, extinction time, flame height, spread rate, and calculating fire intensity. Comparative analysis revealed: (1) Fire intensity ranking: P. armandii needles > A. dealbata leaves > P. armandii branches > A. nepalensis leaves > P. armandii bark > A. dealbata branches > A. nepalensis branches > A. dealbata bark > A. nepalensis bark; (2) The biological firebreaks composed of A. nepalensis and A. dealbata in Yuanbaoshan exhibited effective flame-retardant performance; (3) Coarse woody fuels negatively affected prescribed burning intensity and effectiveness. By quantifying fire behavior differences among these species, this study provides scientific support for fuel management and fire-resistant species selection in central Yunnan, while offering practical guidance for prescribed burning strategies in the Jin Dian Yuanbaoshan Forest Area.

1. Introduction

Forest fires refer to uncontrolled combustion phenomena occurring in forested areas, woodland ecosystems, and their adjacent ecological zones. When flames breach predetermined containment boundaries and escape human control, they not only jeopardize human lives and property but also exert profound and enduring impacts on natural ecosystems, climatic patterns, and socioeconomic development [1]. Against the backdrop of rapid societal advancement and intensifying global warming, forest ecosystems now face increasingly severe challenges. Statistical evidence indicates a global upward trend in the frequency, severity, and duration of wildfires [2]. In 2023, Greece experienced the largest recorded wildfire in EU history; the burned area exceeded 96,000 ha [3]. That same year, Canadian wildfires devastated nearly 7.8 × 10⁶ ha of forestland, emitting roughly 3 × 10⁹ metric tons of CO₂—accounting for 25% of global wildfire carbon emissions [4]. On 8 August 2023, at least three wildfires ignited on Maui Island, Hawaii, USA, scorching over 8.78 ha and causing catastrophic casualties and economic losses [5]. As a nation possessing significant forest resources that constitute vital components of global woodlands, China ranks among the countries with the highest forest fire risks and most demanding fire prevention responsibilities [6]. The southwestern forest region of China exhibits particularly hazardous fire behavior due to its unique geographical setting, complex microclimates, and heavy fuel loads [7]. Serving as an ecological barrier within this region for mitigating desertification and conserving soil and water resources, central Yunnan demonstrates high forest fire susceptibility and risk, characterized by diverse fuel types, variable topography, and a long history of fire use in traditional practices [8].

According to the 10th National Forest Inventory of China, Yunnan Province possesses a forest area of 2.1 × 10⁷ ha, which constitutes 55.25% of the province’s total land area. The standing timber volume is 2.1 × 10⁹ m³ [9]. Forest fires represent a recurrent natural disaster in Yunnan, characterized by widespread occurrence, sudden onset, severe destructiveness, and challenging containment, warranting its designation as a national priority region for forest fire prevention [10]. Located in central Yunnan, this area exhibits high fire frequency within the province. The region features diverse vegetation types and abundant forest resources, dominated by native coniferous species (e.g., Pinus yunnanensis, P. armandii) with extensive distribution of native broadleaf forests (e.g., Quercus variabilis, Alnus nepalensis). The topography primarily consists of plateaus and mountains, with a subtropical plateau monsoon climate exhibiting reduced precipitation during winter and spring. The convergence of climatic conditions, vegetation characteristics, and anthropogenic factors contributes to elevated forest fire risks in central Yunnan [11].

Fire behavior refers to the dynamic physical characteristics manifested during forest fuel combustion, with key parameters including flame spread rate, heat release rate, and burning duration that critically determine fire spread dynamics and suppression difficulty [12]. Based on spatial combustion patterns, forest fires are classified into three fundamental types: surface fires, crown fires, and ground fires. Surface fires burn across the forest floor consuming litter, grasses, and low shrubs. In contrast, ground fires burn beneath the surface in deep layers of organic matter like peat or duff, which can smolder for extended periods. Crown fires propagate through the canopy of trees. When flame height from a surface fire exceeds a critical threshold, it may undergo combustion mode transition, developing into more destructive crown fires that establish a complex, multi-layered fire structure, with spot fires potentially occurring under extreme fire behavior conditions [13]. The JinDian Yuanbaoshan Forest Area, a representative high fire-risk zone in central Yunnan, is characterized by its typical forest plantation primarily composed of Chinese fir and drought-resistant wintergreen, exhibits surface fuel loads significantly exceeding critical thresholds. This results in a very high likelihood that an ignition will develop into a sustained surface fire, which in turn carries a substantial risk of transition into a crown fire due to the abundant ladder fuels and coniferous canopy. Prescribed burning effectively regulates surface fuel structure by reducing fuel loads to safe thresholds—specifically defined as the critical fuel load value below which, even if ignition occurs, both fire intensity and spread rate can be effectively controlled. This significantly reduces the risk of fire escalation (e.g., transition from surface to crown fire) and ensures the safety and feasibility of suppression operations—while significantly suppressing potential fire intensity development. Through standardized laboratory combustion experiments, this study systematically quantified key fuel properties and fire behavior parameters (e.g., flame duration, spread rate, and heat release) of typical tree species fuels in this region. These observed data provide essential input parameters for developing regional fire behavior prediction models. It is important to note that this study constitutes a fundamental, first step. While these parameters are a critical prerequisite, translating them into a fully realized predictive model presents significant challenges that remain to be addressed in future work. These include integrating these inputs with field-scale variables like topography and meteorology, validating the models against actual wildfire behavior, and defining the specific architecture for such a regional framework. Therefore, this paper does not propose a ready-to-use model but rather provides the indispensable empirical foundation for one. The potential usefulness of this foundation is to enable the future development of models that could, in turn, offer a scientific basis for optimizing prescribed burning strategies.

Pinus armandii, a montane cool-temperate coniferous species in western subtropical China, exhibits broad ecological adaptability and vigorous natural regeneration capacity [14]. As a predominant timber species in southwestern China and a common conifer on the Yunnan-Guizhou Plateau, it demonstrates extensive distribution and remarkable cold tolerance. Its needles contain high concentrations of volatile oils. Mature trees often form dense, continuous canopies with low-hanging branches, which facilitates the vertical connectivity of fuel. With the oil, this structure allows surface fires spread upward to the canopy quickly. This fire type, characterized by high-intensity and complete canopy consumption, often leads to significant short to medium-term alterations to forest ecosystem structure and function. These alterations include high tree mortality, loss of seed sources, increased soil erosion, and pronounced changes in habitat and biogeochemical cycles. While fire is a natural component of many ecosystems, the high frequency and intensity of crown fires, particularly under altered climate and fuel conditions, can exceed historical return intervals and resilience thresholds. Due to the potential for these lasting impacts on ecosystem services and protection forestry objectives, mitigating crown fire risk is a management priority in these commercially and ecologically important southwestern forests. Alnus nepalensis (commonly called Nepalese alder), a fast-growing deciduous tree of the Betulaceae family, possesses notable ecological and physiological traits including soil fertility enhancement and strong coppicing regeneration capacity, endowing it with significant restoration potential in degraded ecosystems [15]. As an important timber and ecological service forest species in southwestern China, A. nepalensis demonstrates exceptional fire-resistant properties. Acacia dealbata (silver wattle), an evergreen tree of the Fabaceae family native to Australia, exhibits strong environmental adaptability and rapid growth, providing effective soil/water conservation and windbreak benefits. This species serves as an excellent ecological restoration and shelterbelt tree. Due to traits such as its relatively high moisture content in foliage, slow-decomposing litter that forms a less flammable fuel bed, and a canopy structure that may moderate understory microclimate, it is frequently employed as a fire-resistant species in reforestation and greenbelt projects in Yunnan Province [16].

2. Materials and Methods

2.1. Study Area

The study area was located in the Jin Dian Yuanbaoshan Forest (25°07′12″~25°08′33″ N, 102°45′36″~102°47′02″ E) in Kunming, Yunnan Province, China, as shown in Figure 1. This forest features a northwest-high, southeast-low topography with local slopes exceeding 40° and an intricate network of V-shaped valleys and short tributaries. The forest spans altitudes of 1920~2350 m, covering a total area of 217.2 ha. Situated at the edge of the central subtropical zone, the area experiences a plateau monsoon climate characterized by distinct wet and dry seasons. Seasonal precipitation is unevenly distributed, leading to a significant increase in forest fire risk during the dry season. The region has a mean annual temperature of 14.8 °C, annual precipitation of 950~1050 mm, and annual evaporation of 1800~2000 mm, 80%–90% of the precipitation occurs during the rainy season from May to September, resulting in notable water deficits during dry periods, causing a significant increase in forest fire risk during the dry season. The average annual wind speed is 2.5 m/s, reaching 4 m/s during spring [17]. This seasonal peak in wind speed coincides with the annual dry period, characterized by low humidity and dormant, flammable vegetation. These sustained spring winds effectively dry fine fuels, facilitate the rapid spread of any ignition, and increase the likelihood of surface fires transitioning into more dangerous crown fires. The convergence of these climatic conditions during spring is a key determinant of the critical fire season, thereby substantially elevating overall fire risk and behavior intensity during this period.

The Jindian Yuanbaoshan Forest Area in Kunming has a high forest coverage rate of 86.3 percent. The main vegetation type is subtropical evergreen broad-leaved forest, and the dominant tree species include Cyclobalanopsis glaucoides, Pinus yunnanensis, and Pinus armandii. Estimated based on the typical sampling method and the stratified random method, in this forest area, the plant community is dominated by the tree layer, which accounts for 46.27 percent, while the shrub layer is severely lacking, constituting only 11.19 percent. The community appearance is primarily characterized by coniferous-broadleaf mixed forests, making up 40.0 percent. The vertical structure exhibits distinct stratification, with coniferous species (such as Pinus yunnanensis and Pinus armandii) occupying the upper canopy and broad-leaved species distributed in the middle and lower layers. The forest resources are abundant [18]. Pure stands of coniferous trees like Pinus yunnanensis and Pinus armandii are distributed in continuous patches, with large surface fuel loads, making them highly susceptible to crown fires. This area is one of the typical high-risk regions for crown fires in China. In the future, prescribed burning could be implemented for fuel management in this region to maintain potential surface fire intensity below 500 kW/m under moderate weather conditions, thereby preventing crown fire initiation by reducing both surface fuel loads and ladder fuel connectivity and significantly reducing the probability of high-intensity fire occurrences.

The investigation employed systematic sampling across representative forest types (Pinus armandii pure stands and coniferous-broadleaf mixed forests). A total of twenty-eight 10 m × 10 m plots were established at uniform 10-m intervals. This design intentionally samples the homogeneous interior forest conditions while avoiding edge zones—thereby reducing errors caused by the complex microclimate and heterogeneous fuel characteristics in edge areas. Consequently, our data accurately represent the fire behavior potential of core forest stands but do not capture the specific fuel dynamics or elevated fire risk potentially associated with forest edges. As illustrated in Figure 2, five 1 m × 1 m quadrats were systematically arranged along diagonal transects within each plot.

2.2. Sample Preparation

This study selected two broadleaf species (Alnus nepalensis and Acacia dealbata) and one coniferous species (Pinus armandii) from the Jin-Dian Yuanbaoshan Forest area. A. nepalensis and A. dealbata are commonly used as fire-resistant species in this region, with A. dealbata being considered a low-fire-risk exotic species. P. armandii, as a dominant tree species in the area, accumulates a significant fuel load with properties that frequently leads to high-intensity crown fires once an ignition occurs. The three tree species differ in their phylogenetic origins, leaf morphology, fire-related adaptations (e.g., Pinus armandii is highly flammable, whereas the two broadleaf species exhibit greater fire resistance), as well as ecological roles and regeneration strategies. These species are widely distributed in central Yunnan and provide comparative significance, ensuring scientific rigor in sample selection. Investigating the flammability and potential fire behavior of these three key species provides critical scientific basis for fuel management and fire-resistant species selection in central Yunnan. Field data collection was completed in May 2025 during Yunnan’s fire season, which extends from 1 December to 15 June. This period is characterized by surface fuel moisture content values typically ranging from 12 to 18 percent in fine fuels (e.g., litter, grasses, and twigs <0.6 cm in diameter), representing conditions conducive to typical combustion behavior. All three selected tree species were sampled between 9:00 and 11:00 AM under similar growth conditions; their leaves, bark, and whole branches were collected and stored in sealed bags. For surface litter sampling, five 1 m × 1 m subplots were set within each sampling site, and the collected litter mainly consisted of dead leaves, dead branches (with a diameter of <0.6 cm), and bark fragments. Each sample was individually collected according to a standard quantity of 5 kg by weight. After collection, each sample was clearly categorized as: needles, branches, and bark of Pinus armandii; leaves, branches, and bark of Alnus nepalensis and Acacia dealbata; as well as surface litter of Pinus armandii, Alnus nepalensis, Acacia dealbata, and Pinus yunnanensis. The surface litter of Pinus yunnanensis includes not only needles but also combustible materials such as dead branches and bark. To ensure better ignition material for the experiment without affecting the data, dead branches and bark were removed. Samples of the same type collected randomly from different plots were mixed, labeled, and transported to the laboratory for subsequent experiments. The primary objective of this laboratory study was to isolate and quantify the fundamental flammability parameters of different fuel types under controlled and reproducible conditions. To achieve this, the spatial arrangement of fuels was maintained from field collection to laboratory setup, ensuring consistency in packing ratio and homogeneity of material distribution.

2.3. Laboratory Experiment

In this experiment, a randomly selected portion of the samples were dried in an electric thermostatic drying oven (XIUILAB, Shanghai, China) at 105 ± 5 °C for 12 h until absolutely dry, to calculate their moisture content. To simulate real forest fire scenarios, the samples used in the combustion experiments were not pre-dried. This study employed a downhill fire spread simulation experiment method. On the adjustable slope combustion bed shown in Figure 3 within a standard combustion laboratory, wind-free conditions were strictly controlled to simulate the fire behavior characteristics of prescribed burning. The combustion bed is 2.03 m long, 1.27 m wide, and 0.34 m high. A 2 cm thick gypsum board was laid at the bottom to prevent heat loss. The slope was set to 15°, and a 1 m × 1 m area was demarcated inside the combustion bed to simulate the field plot range.

Prior to each experiment, sample mass was measured using an electronic balance (precision: 0.01 g) before uniform distribution within the 1 m × 1 m test area, with ignition achieved using a flame torch applied from the platform’s upper edge downward to simulate downhill fire spread. Combustion duration (ignition-to-extinction) and smoldering time were recorded with a stopwatch, while spread rate was calculated through temporal analysis. Non-contact infrared thermometry (DTM-T1, DELIXI ELECTRIC, Shanghai, China) measured combustion and smoldering temperatures. At the same time, the height of the flame was measured using a steel tape measure that was perpendicular to the surface of the combustion bed. Spatial temperature profiling employed nine uniformly distributed K-type thermocouples (80 µm diameter) acquiring data at 3.6 s intervals throughout combustion. Post-test mass quantification using the electronic balance enabled fuel consumption rate calculations.

2.3.1. Determination of FMC

Fuel moisture content (FMC), defined as the percentage of water mass to the dry mass of combustible material, serves as a critical indicator of fuel wetness and a key determinant of forest fire risk and fire behavior [19]. FMC directly influences the ignition probability of forest fires and indirectly affects fire intensity and spread rate, representing an essential factor in assessing fuel flammability. The FMC is calculated as follows:

$$\mathrm{F}\mathrm{M}\mathrm{C}={\displaystyle \frac{{\mathrm{m}}_0-\mathrm{m}}{{\mathrm{m}}_0}}\times 100\%$$

(1)

where m₀ represents the fresh weight (g) and m denotes the oven-dry weight (g).

2.3.2. Calculation of Fire Intensity

Fire intensity, a key physical quantity characterizing the energy release rate per unit fireline length, is commonly used to quantify the destructive force and spread potential of wildfires. As a critical indicator of forest fire combustion severity, it significantly influences fire development, ecological impacts, and suppression difficulty. The fire intensity I (unit: kW/m) can be estimated from flame height [20] as follows:

$$\mathrm{I}=273{\mathrm{h}}^{2.17}$$

(2)

where h represents the flame height (m).

2.3.3. Determination of Burn-Up Rate

In forest fires, the mass loss rate refers to the proportion of combustible material that actually participates in the combustion reaction under specific fire conditions, serving as a crucial indicator for assessing the damage level to forest ecosystems. This parameter quantitatively evaluates fire-induced impacts on vegetation, soil, and ecological systems, with its value directly determining the total energy released during combustion [21]. Research demonstrates a significant negative correlation between mass loss and both wind speed and terrain slope: increased wind velocity or steeper slopes accelerate flame spread rates, thereby reducing fire residence time per unit area and consequently decreasing combustion completeness and mass loss rate [22]. Calculated as the ratio of consumed fuel mass to initial fuel load, the mass loss rate represents a key parameter characterizing forest fuel combustion efficiency, providing quantitative measurement of actual fuel consumption during fires and serving as essential evidence for evaluating fire intensity and combustion completeness. The calculation formula is as follows:

$$\mathrm{R}={\displaystyle \frac{{\mathrm{m}}_1-{\mathrm{m}}_2}{{\mathrm{m}}_1}}\times 100\%$$

(3)

where m₁ denotes the pre-combustion fuel mass (kg) and m₂ represents the post-combustion residual fuel mass (kg).

2.3.4. Statistical Analysis

Statistical analyses including one-way ANOVA and arithmetic mean calculations were performed using SPSS 26 (SPSS Institute, Inc., Chicago, IL, USA) to process both field-collected fuel characteristic data and experimentally measured potential fire behavior parameters, with all statistical results visualized using Origin 202 (OriginLab Corporation, Northampton, MA, USA).

3. Results

3.1. The Potential Fire Behavior Characteristics of Three Tree Species

Potential fire behavior refers to the comprehensive characterization of the combustion characteristics and dynamic development processes that combustible materials may exhibit under specific laboratory conditions. It encompasses the entire dynamic sequence from ignition, flame propagation to eventual extinction [23]. Analysis of potential fire behavior in specific forest areas enables assessment of post-ignition fire spread probability and probable development trends, while also identifying optimal windows for prescribed burning to ensure controllable fire behavior and achieve desired outcomes [24]. As an important foundation for the scientific management of forest fires, understanding potential fire behavior can provide key support for optimizing prescribed burning from empirical practice to science-based management operations, thereby enhancing the safety, precision, and ecological benefits of such burning activities [25]. The potential fire behavior characteristics of Pinus armandii, Alnus nepalensis, and Acacia dealbata are presented in

Table 1. Potential fire behavior characteristics of Pinus armandii. (The “Moisture content” in this table is the measurement obtained after drying a portion of the samples in an oven for 12 h; The symbol “±” indicates the standard deviation; The “Base” was collected from the surface litter of Yunnan pine within the study area and used as the material for ignition samples. The “Mass ratio” refers to the ratio of the weight of the base and the added experimental sample.).

Mass Ratio	Component	Moisture Content (%)	Burning Duration (s)	Smoldering Duration (s)	Combustion Temperature (°C)	Smoldering Temperature (°C)	Flame Height (cm)	Spread Rate (m·min−1)	Mass Loss Rate (%)	Fire Intensity (kW·m−1)
3:1	Base + Pine needles	14.13	293	90	711.4 ± 48.9	324.3 ± 77.5	60 ± 5	0.21	81.60	90.11
			301	106	633.6 ± 52.5	320.8 ± 71.5	53 ± 8	0.19	76.40	68.84
			290	97	653.2 ± 47.3	306.8 ± 61.9	57 ± 6	0.21	79.00	80.61
3:1	Base + Branch	12.11	365	108	672.5 ± 44.8	306.2 ± 37.4	47 ± 3	0.16	65.55	53.04
			349	80	689.5 ± 47.5	285.7 ± 41.2	44 ± 8	0.17	70.30	45.97
			375	111	666.6 ± 42.9	305.6 ± 37.5	38 ± 7	0.16	73.11	33.44
3:1	Base + Bark	12.90	418	154	642.1 ± 72.6	281.8 ± 22.7	37 ± 3	0.14	63.32	31.56
			343	141	661.5 ± 68.5	290.8 ± 25.4	37 ± 4	0.18	69.43	31.56
			400	145	651.7 ± 61.5	283.5 ± 24.7	39 ± 4	0.15	68.33	35.38
5:1:1:1	Base + Pine needles + Branches + Bark	13.05	391	142	622.0 ± 23.8	251.5 ± 46.1	36 ± 7	0.15	60.62	29.74
			384	137	603.8 ± 21.7	237.6 ± 39.6	33 ± 6	0.16	57.33	24.62
			398	146	625.4 ± 26.9	263.3 ± 48.2	38 ± 3	0.15	66.56	33.44

,

Table 2. Potential fire behavior characteristics of Alnus nepalensis.

Mass Ratio	Component	Moisture Content (%)	Burning Duration (s)	Smoldering Duration (s)	Combustion Temperature (°C)	Smoldering Temperature (°C)	Flame Height (cm)	Spread Rate (m·min−1)	Mass Loss Rate (%)	Fire Intensity (kW·m−1)
3:1	Base + Leaves	21.98	308	104	587.9 ± 27.5	294.5 ± 74.2	42 ± 8	0.19	63.60	41.55
			324	116	576.0 ± 33.3	272.8 ± 43.2	45 ± 7	0.18	66.80	48.27
			310	113	584.5 ± 21.3	297.5 ± 50.5	38 ± 6	0.19	70.30	33.44
3:1	Base + Branch	16.10	426	110	587.0 ± 25.7	298.7 ± 28.7	28 ± 4	0.14	50.18	17.24
			390	95	608.5 ± 24.2	313.9 ± 33.7	27 ± 3	0.15	61.99	15.93
			405	105	589.1 ± 28.2	268.0 ± 31.2	31 ± 4	0.15	63.75	21.50
3:1	Base + Bark	18.28	428	61	610.1 ± 41.9	279.2 ± 30.5	27 ± 4	0.14	50.35	15.93
			400	60	624.5 ± 32.5	280.5 ± 23.2	22 ± 3	0.14	57.04	11.25
			436	71	585.0 ± 30.1	310.5 ± 23.4	24 ± 4	0.15	57.07	12.34
5:1:1:1	Base + Leaves + Branches + Bark	19.38	480	83	558.2 ± 32.6	296.5 ± 22.0	26 ± 3	0.13	37.05	14.67
			456	76	527.3 ± 28.8	257.9 ± 24.7	18 ± 5	0.13	32.33	6.61
			494	107	596.1 ± 38.3	307.6 ± 25.8	30 ± 3	0.12	45.58	20.02
3:1	Base + surface litter (control group).	15.43	436	115	624.2 ± 57.7	214.5 ± 20.2	25 ± 5	0.14	56.50	13.48
			455	127	637.7 ± 52.6	226.7 ± 18.6	33 ± 6	0.13	59.79	24.62
			424	107	588.1 ± 48.3	209.5 ± 21.7	23 ± 7	0.14	53.07	11.24

, and

Table 3. Potential fire behavior characteristics of Acacia dealbata.

Mass Ratio	Component	Moisture Content (%)	Burning Duration (s)	Smoldering Duration (s)	Combustion Temperature (°C)	Smoldering Temperature (°C)	Flame Height (cm)	Spread Rate (m·min−1)	Mass Loss Rate (%)	Fire Intensity (kW·m−1)
3:1	Base + Leaves	22.05	312	90	605.1 ± 37.5	316.6 ± 25.5	51 ± 4	0.19	74.23	63.33
			348	107	571.7 ± 43.3	282.7 ± 28.7	48 ± 5	0.17	71.59	55.52
			305	77	621.7 ± 47.1	304.5 ± 25.2	50 ± 3	0.20	78.73	60.66
3:1	Base + Branch	18.35	500	117	594.5 ± 32.3	283.2 ± 34.9	34 ± 6	0.12	72.17	26.24
			473	105	597.4 ± 28.1	275.7 ± 36.9	42 ± 7	0.13	62.97	41.55
			406	96	626.8 ± 34.4	267.4 ± 32.6	39 ± 6	0.15	70.20	35.38
3:1	Base + Bark	17.83	513	177	578.1 ± 46.1	256.6 ± 42.4	27 ± 5	0.12	57.25	15.93
			441	145	600.0 ± 48.8	285.1 ± 32.7	26 ± 6	0.14	50.06	14.68
			431	133	601.4 ± 35.3	266.5 ± 43.2	23 ± 8	0.14	54.89	10.21
5:1:1:1	Base + Leaves + Branches + Bark	18.81	477	75	614.3 ± 72.7	270.4 ± 65.9	34 ± 3	0.13	42.69	26.27
			489	94	626.9 ± 69.2	288.3 ± 67.0	35 ± 6	0.12	45.53	27.98
			483	101	617.6 ± 73.5	265.5 ± 61.2	38 ± 3	0.12	43.85	33.44

, respectively.

3.1.1. Combustion Duration and Temperature

Combustion duration refers to the sustained burning period from initial ignition to complete flame extinction [26], while combustion temperature represents the peak temperature attained during heat release, serving as a critical parameter for assessing fire intensity and destructive potential [27]. Comparative analysis of the three species revealed significant differences in these parameters: Pinus armandii needles exhibited the highest combustion temperature (711.4 °C) and bark showed the longest duration (400 s), attributable to their low moisture content and high resin concentration. In contrast, Alnus nepalensis and Acacia dealbata demonstrated lower temperatures (558.2~626.8 °C) but generally prolonged combustion, reflecting the persistent burning characteristics of fire-resistant broadleaf species. Notable thermal heterogeneity was observed in P. armandii components (needles vs. bark), whereas A. nepalensis and A. dealbata exhibited minimal intra-species temperature variation. Table 1, Table 2 and Table 3 consistently demonstrate that fine fuels (needles/leaves) ignite more readily than coarse fuels (branches/bark).

3.1.2. Smoldering Duration and Temperature

Smoldering duration refers to the persistent glowing combustion period after flame extinction until complete extinguishment [28], representing a slow oxidation process within porous media characterized by flameless combustion, low temperatures, and continuous release of combustible gases [29]. This phenomenon commonly occurs in loose organic layers or high-porosity plant tissues within forest ecosystems, exhibiting significantly lower combustion rates than flaming fires but potentially lasting for hours or even days. In this study, Pinus armandii needles demonstrated the longest smoldering duration (154 s) and highest temperature (324.3 °C), where prolonged high-temperature smoldering substantially increases reignition risks and fire suppression challenges post-flame; Alnus nepalensis mixed components showed variable smoldering durations (83~177 s) attributable to fuel heterogeneity-induced instability. Fuel heterogeneity refers to the non-uniformity and diversity of combustible materials in terms of their physical properties, chemical composition, or spatial distribution. These differences significantly influence fire behavior (such as combustion rate, intensity, and spread patterns), thereby leading to instability in burning processes [30]; whereas Acacia dealbata bark, owing to its dense structure, exhibited the shortest smoldering duration (60 s) and lowest temperature (279.2 °C).

3.1.3. Flame Height

According to Table 1, the average flame height of Pinus armandii ranges from 36 to 60 cm. Table 2 shows that the average flame height of Alnus nepalensis ranges from 22 to 45 cm, while Table 3 indicates that the average flame height of Acacia dealbata ranges from 23 to 51 cm. Among these, the needles of Pinus armandii exhibit the highest average flame height at 60 cm, whereas the flame heights of Alnus nepalensis are generally lower, with its bark showing the lowest average flame height of 22 cm. In the Acacia dealbata group, the leaves demonstrate significantly higher flame heights compared to other components (Figure 4), indicating that leaf morphology and chemical composition have a notable influence on flame propagation. The ranking of flame heights is: Pinus armandii > Acacia dealbata > Alnus nepalensis. This order reflects the suppressive effect of the dense structure of broadleaf species on flame development, which is negatively correlated with moisture content.

3.1.4. Spread Rate

Fire spread rate, defined as the perpendicular advancement distance of the fire front per unit time, represents a critical parameter determining fire development dynamics, suppression difficulty, and ecological impacts [31]. When spread rates remain below operational response thresholds (<2 km/h), direct suppression methods become feasible, whereas exceeding critical rates necessitates indirect containment strategies like firebreak construction. This study quantified linear spread rates as the ratio between flame front advancement distance (ignition-to-extinction) and corresponding time intervals. Experimental results showed Pinus armandii with spread rates of 0.14~0.21 m/min (Figure 5), Alnus nepalensis demonstrating slower propagation (0.11~0.19 m/min, with bark at minimum 0.11 m/min), and Acacia dealbata exhibiting intermediate values (0.12~0.20 m/min), all characteristic of slow surface fires (<2 m/min). The three species displayed comparable downhill spread patterns under the 15° slope configuration, exhibiting uniform combustion characteristics including low flame heights and consistent spread behavior typical of gentle terrain conditions.

3.1.5. Mass Loss Rate

The mass loss rate, defined as the percentage of consumed fuel mass relative to the initial mass, serves as a fundamental parameter for quantifying combustion efficiency [21]. Experimental results demonstrated mass loss rates of 60.62%~81.6% for Pinus armandii, 37.5%~70.3% for Alnus nepalensis, and 42.69%~78.73% for Acacia dealbata, with the species-level hierarchy (P. armandii > A. dealbata > A. nepalensis) showing a negative correlation with fuel moisture content.

Figure 6 clearly demonstrates that control group samples (substrate with Alnus nepalensis surface litter) containing leaves, branches (<0.6 cm diameter) and bark exhibited higher mass loss rates compared to mixed samples of A. nepalensis leaves, branches and bark, potentially attributable to differences in physiological status and moisture content between the fuel types. Dead fuels showed significantly greater mass loss than live fuels, indicating their superior capacity to meet sustained combustion requirements under identical fire conditions through more efficient heat accumulation that promotes complete combustion, whereas the moisture content in live fuels absorbs substantial thermal energy, thereby inhibiting combustion efficiency and reducing mass loss rates. Samples containing branches and bark displayed lower mass loss compared to foliar-only samples, confirming that larger particulate fuels (branches/bark) physically impede combustion progression through reduced surface-area-to-volume ratios and modified heat transfer dynamics.

3.1.6. Fire Intensity

Fire intensity quantifies the heat energy released per unit time along a unit length of fire front and serves as a fundamental parameter in wildfire behavior research [32]. Experimental results demonstrated species-specific intensity ranges with Pinus armandii exhibiting the highest values from 29.74 to 90.11 kW/m, Alnus nepalensis showing consistently lower intensities between 11.25 and 48.27 kW/m, and Acacia dealbata displaying intermediate values ranging from 10.21 to 63.33 kW/m (Figure 7), all measurements classifying as low-intensity fires according to the sub-750 kW/m threshold [33]. The observed variation in fire intensity primarily reflects differences in fuel properties including volatile content, packing ratio, and moisture status among the three species.

3.2. Thermocouple Temperature Data

In this experiment, thermocouples positioned 10 cm above the burning fuels recorded air temperature variations during the initial 300-s combustion period. Figure 8a,c,e illustrates the combustion state of leaf samples mixed with substrate after 150 s, while Figure 8b,d,f presents the corresponding thermocouple temperature profiles. The downhill fire spread exhibited relatively low propagation rates, with fire fronts developing uniform horizontal advancement patterns under gravitational influence, forming continuous and stable propagation zones along the slope. Maximum recorded temperatures during leaf combustion reached 594.9 °C for Pinus armandii, 516.5 °C for Acacia dealbata, and 528.3 °C for Alnus nepalensis, establishing the thermal hierarchy of P. armandii > A. dealbata > A. nepalensis. During the preheating phase, significant thermal energy was consumed for moisture evaporation, with P. armandii needles containing higher concentrations of resinous and volatile compounds coupled with lower moisture content enabling rapid temperature escalation, whereas the greater moisture content in A. nepalensis and A. dealbata foliage required more energy for vaporization, consequently reducing the available energy for flame temperature elevation and resulting in lower combustion temperatures. The experimental data demonstrate clear correlations between fuel properties, thermal energy distribution, and combustion performance across the three species.

The simulated experiments involving substrate mixtures with branches, bark, and combined leaf-branch-bark samples produced temperature profiles recorded by thermocouples as shown in Figure 9a–i. Maximum temperatures measured during branch combustion reached 544.1 °C for Pinus armandii, 468.4 °C for Alnus nepalensis, and 486.5 °C for Acacia dealbata, establishing the thermal performance hierarchy of P. armandii > A. dealbata > A. nepalensis for woody fuels. Bark combustion exhibited lower peak temperatures at 504.6 °C, 433.1 °C, and 453.3 °C respectively for the three species, maintaining the same ranking pattern. Combined leaf-branch-bark mixtures demonstrated intermediate maximum temperatures of 564.2 °C, 479.5 °C, and 499.4 °C, again following the established species order. Comparative analysis revealed consistent suppression effects from bark and branch components, with A. nepalensis showing the strongest fire inhibition characteristics, followed by A. dealbata, while P. armandii displayed the least suppression capacity, indicating superior fire resistance properties in A. nepalensis woody components compared to the other species. The temperature differentials between pure foliar and mixed component combustion further confirm the modifying effects of coarser fuels on overall fire behavior through altered heat transfer dynamics and combustion efficiency.

3.3. Effects of Fuel Morphology on Fire Behavior

Fuel morphology directly determines combustion characteristics, consequently influencing fire behavior intensity and propagation patterns [34]. Analysis of potential fire behavior across different fuel types revealed distinct patterns: fine fuels (needles/leaves) from Pinus armandii, Alnus nepalensis, and Acacia dealbata exhibited rapid combustion with relatively brief flaming durations, establishing clear morphological controls on combustion dynamics through variations in surface-to-volume ratios, heat transfer efficiency, and fuel continuity. The experimental data systematically demonstrate how fuel structural properties modulate energy release rates and combustion completeness across different size classes and species.

The flame height, fire intensity and burn rate of the fine combustibles of the three tree species are all higher than those of the branch and bark combustion. The flaming combustion time of the medium combustibles is between that of the fine combustibles and the coarse combustibles, which will prolong the duration of the fire line and have a transitional effect on the spread of the fire. The flaming combustion duration of the coarse combustibles is the longest, while the flame height, fire intensity and burn rate are the lowest. By comparing the fire behavior of the combustion of mixed samples of leaves, branches and bark, it was found that the coarse combustibles have a significant inhibitory effect on the fire behavior. Before planned burning, it is necessary to remove the coarse combustibles that are likely to affect the fire behavior in advance to avoid the fire not reaching the fire intensity required for resource management during the planned burning and to reduce the effect of the planned burning on the management of surface combustibles.

4. Discussion

4.1. Stand Characteristics of Jin-Dian Yuanbaoshan Forest Area

The Pinus armandii pure stands exhibit a relatively high canopy cover of 0.7 that significantly restricts understory vegetation development, resulting in sparse herbaceous growth, while the homogeneous fuel composition primarily consisting of litter, fallen dead branches, and bark fragments creates typical surface fire conditions, further facilitated by the low average crown base height of 1.6 m maintaining good vertical fuel continuity [35]. The Alnus nepalensis stands contain abundant Urtica dioica ground cover that actively grows during its March–September growing period, which coincides with the fire season with their straight trunks and high branching preventing crown fire formation as leaves and branches shed early in the fire season [36]. Acacia dealbata demonstrates strong drought tolerance and environmental adaptability with vigorous root suckering, rapid growth, and high canopy density that suppress understory vegetation, making it well-adapted to Kunming conditions and a preferred species for biological firebreaks [37], though its aggressive expansion in Yuanbaoshan requires selective management to retain superior phenotypes with enhanced fire resilience, optimal canopy structure, and controlled spread, while its dense canopy naturally reduces surface fuels and inhibits fire spread.

Experimental data suggest that during the fire season, the low moisture content and high fuel loads in Yuanbaoshan Forest create favorable conditions for ignition and combustion [38]. In P. armandii pure stands, surface fires could readily transition to crown fires due to vertical fuel continuity, necessitating proactive fuel management to mitigate high-intensity wildfire risks. In contrast, A. nepalensis stands benefit from fire-suppressive effects of Urtica dioica ground cover during its growing season when high moisture content slows fire spread, while A. dealbata stands’ sparse understory vegetation and limited surface fuels inherently restrict fire intensity. This study elucidates species-specific fire behavior differences in Yuanbaoshan to identify high-risk factors. Our findings, which map these risks to specific locations, are directly applicable to optimizing fuel management. This allows managers to prioritize areas for different treatments—such as targeted fuel reduction in high-risk Pinus armandii stands or the strategic expansion of Acacia dealbata firebreaks—thereby developing more effective and science-based fire prevention plans.

4.2. Variation in Potential Fire Behavior Among Tree Species

Simulated combustion experiments in Yuanbaoshan Forest’s Pinus armandii pure stands and firebreak belts revealed significantly lower moisture content, higher fire intensity, longer flaming duration, and elevated temperatures in P. armandii stands, compared to firebreak belts dominated by Alnus nepalensis and Acacia dealbata, which exhibited higher moisture content, lower fire intensity, and reduced flaming temperatures. The combustion intensity hierarchy demonstrated P. armandii needles > A. dealbata leaves > P. armandii branches > A. nepalensis leaves > P. armandii bark > A. dealbata branches > A. nepalensis branches > A. dealbata bark > A. nepalensis bark. It demonstrates that both Alnus nepalensis and Acacia dealbata significantly inhibit the spread of forest fires, with the bark of Alnus nepalensis exhibiting the most pronounced inhibitory effect, consistent with findings from Li Shiyou [39], Su Wenjing [40], and Zhang Yunsheng [41]. These results confirm the effective reduction of fire spread rates and intensity in Yuanbaoshan’s A. nepalensis and A. dealbata firebreaks, while comparative analysis of their potential fire behavior remains limited. Experimental data show A. nepalensis’s significantly lower values than A. dealbata across critical parameters, including combustion temperature, fire intensity, and flame height, suggesting superior fire resistance and greater potential for firebreak applications.

4.3. Critical Factors Affecting Prescribed Burning Efficacy

Prescribed burning refers to the controlled application of fire under specific environmental conditions to wildland fuels in their natural state or after artificial modification, with its core objective being to confine the fire within a predetermined area while achieving the fire intensity and spread rate required for resource management. Prescribed burning is influenced by factors such as weather conditions, the physical state of fuels, and topography. Conducting prescribed burning requires selecting days with a fire danger rating of Level 3 or below and wind force of Level 3 or below, and determining the ignition time based on factors such as forest conditions, air humidity, temperature, and fuel moisture content [42]. This study specifically examined fuel physical state impacts on prescribed burning and downhill fire characteristics at 15° slopes, demonstrating that larger fuel particles with higher moisture content fail to achieve required fire intensities for effective fuel management, while fires on 15° slopes exhibited low spread rates, prolonged burning durations, and uniform complete combustion, consistent with typical low-gradient fire behavior, aligning with Xu Hongsheng’s findings [43]. Although our research investigated fire behavior differences among Pinus armandii, Alnus nepalensis, and Acacia dealbata, along with fuel physical state effects on prescribed burning, the study’s limitations include using only 15° slope simulations for these three species, with future research planned to measure heat values, flaming radiation, and varied slope angles/ignition methods, to better understand these species’ potential fire behavior and slope impacts on prescribed burning efficacy.

5. Conclusions

This study provides a comprehensive comparative analysis of fire behavior characteristics in Pinus armandii, Alnus nepalensis, and Acacia dealbata forests, with particular focus on their implications for prescribed burning operations, revealing three key findings. First, Pinus armandii stands in Jin-Dian Yuanbaoshan Forest exhibited critically low moisture levels, ranging 12.90%~14.13% during May in Kunming, indicating high flammability and elevated fire risk, whereas Alnus nepalensis and Acacia dealbata stands demonstrated higher moisture content between 15.43%~22.05%, resulting in extended flaming durations that effectively retard fire spread and provide measurable fire suppression benefits. Second, the biological firebreaks established with Alnus nepalensis and Acacia dealbata in Yuanbaoshan prove particularly valuable for wildfire management, with Alnus nepalensis demonstrating superior fire resistance characteristics compared to Acacia dealbata. Third, coarse woody fuels significantly inhibit fire behavior and frequently prevent prescribed burns from achieving necessary intensity thresholds, mandating pretreatment removal of such materials prior to prescribed burning operations in Yuanbaoshan to ensure management objectives are met. Collectively, these findings underscore the essential role of strategic fuel management in preventing surface-to-crown fire transitions during prescribed burning activities in Jin-Dian Yuanbaoshan Forest.