Next Article in Journal
Regional-Scale Assessment of Burn Scar Mapping in Southwestern Amazonia Using Burned Area Products and CBERS/WFI Data Cubes
Previous Article in Journal
The Interacting Influence of Fire and Tree Characteristics on Douglas-Fir Beetle Host-Tree Selection Five Years Post-Fire
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Fire
Volume 7
Issue 3
10.3390/fire7030065
fire-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Ignition of Forest Fires by Cigarette Butts: Using Pinus massoniana Needles as an Example

by
Yunlin Zhang
1,2,* and
Lingling Tian
1
1
School of Biology Sciences, Guizhou Education University, Gaoxin St. 115, Guiyang 550018, China
2
Key Laboratory of Ecology and Management on Forest Fire in Universities of Guizhou Province, Guizhou Education University, Gaoxin St. 115, Guiyang 550018, China
*
Author to whom correspondence should be addressed.
Fire 2024, 7(3), 65; https://0-doi-org.brum.beds.ac.uk/10.3390/fire7030065
Submission received: 20 January 2024 / Revised: 19 February 2024 / Accepted: 23 February 2024 / Published: 24 February 2024
(This article belongs to the Section Fire Science Models, Remote Sensing, and Data)
Browse Figures
Review Reports Versions Notes

Abstract

:
As a cigarette butt falls onto the forest surface fuel, it first smolders the fuel, then ignites into flames, and spreads as forest fire under certain conditions. In this study, the needles under a typical stand of P. massoniana were used as the research object. Needle beds with different moisture content and packing ratios were constructed indoors. Cigarette butt-ignition experiments were conducted under different wind velocities, and 30 experiment cycles were conducted under different conditions. There was a total of 5 (packing ratio) × 4 (moisture content) × 6 (wind velocity) = 120 sets of conditions, and a total of 3600 ignition experiments were conducted. The results showed that (1) the total ignition probability of the cigarette butts was 2.36%, which only occurred when the fuelbed moisture content was <10% and the wind velocity was >1 m/s. The ignition time of cigarette butts ranged from 2.73 to 7.25 min. (2) The fuelbed moisture content and wind velocity significantly influenced the ignition probability and time. With an increase in moisture content, the ignition probability of cigarette butts decreased, while the time required for ignition showed an increasing trend. Wind velocity had a dual effect on ignition. The ignition effect was optimal at a wind velocity of 4 m/s. With an increase in wind velocity, the ignition probability first increased and then decreased, and the ignition time first decreased and then increased. (3) The packing ratio had no significant effect on the ignition probability; however, the ignition time significantly decreased as the packing ratio increased. (4) The logistic regression method (LRM), general linear method (GLM), and nonlinear regression method (NLM) were used to establish a prediction model of ignition probability. The prediction effect of GLM was the worst, followed by LRM, and the NLM had the best prediction effect. The GLM was selected to establish the ignition time model, and the error was also within the allowance range. This study elucidated the underlying mechanism of factors affecting cigarette butt-based fuel ignition. In addition, the established prediction model provides a reference for human-caused forest fires and is highly significant for forest fire prevention.

1. Introduction

Guizhou Province is located in southwestern China, mostly in the agricultural–forestry transitional zone with high mountains and steep slopes [1]. Forest fires are extremely difficult to extinguish in such areas and pose a serious threat to the safety of people, infrastructure, social stability, and the environment [2,3]. According to the reported statistics, 18496 forest fires occurred in Guizhou Province between 2001 and 2017, of which over 95% were human-caused [4]. Therefore, controlling man-caused fires is a top priority for forest fire prevention and extinguishment in Guizhou Province [4]. A man-caused fire is a forest fire initiated by a human source, such as unattended campfires, rubbish burning, and carelessly discarded cigarettes. Specific anthropogenic fire sources are generally influenced by meteorological conditions and the physical and chemical properties of fuel [5,6,7,8,9,10]. Analyzing the mechanism of human-induced ignition of forest fuels in Guizhou Province and establishing a prediction model for ignition probability and time are significant for the study of forest fire occurrence prediction, prevention, and extinguishment in Guizhou Province and also provide a reference for other man-caused fire research.
Fire sources in Guizhou Province include sacrificial fires, cigarette butts, playing with fire, burning waste, and refining mountains. Cigarette butts, which are a special source of man-caused fire, account for approximately 30–40% of the fire sources generated during the human life process in the study area [4]. Since cigarette butts thrown onto forest fuel initially smolder, forest fire monitoring systems cannot easily detect them. These fires reach a disaster scale by the time of detection, missing the best control and extinguishment time [11,12]. Research has shown that the probability and time of cigarette butt ignition are not fixed and unchanging but are influenced by the fuel type, fuelbed structure, physicochemical properties, and meteorological conditions [13,14,15]. However, there are currently few studies on cigarette butts igniting forest fires [15]. Therefore, determining the ignition probability and time of cigarette butts can promote the reduction and control of cigarette-led forest fires and also provide a reference for other fire source research.
Although cigarette butts are one of the main causes of forest fires, only a few ignition experiments or analyses of cigarette butts have been conducted. Markalas et al. used cigarette butts to ignite pine needle beds with a 5% moisture content and found that cigarette butts could not ignite needle beds under no-wind conditions, and the ignition probability significantly increased with increasing wind velocity [16]. Satoh et al. found that when the wind velocity exceeded a certain threshold, the cigarette butts would be blown away or extinguished, making fuel ignition impossible [17]. However, when the wind velocity was low, increasing the wind speed promoted the fuel ignition by a cigarette butt. Sun et al. studied the litter under a Querus mongolian forest, a typical broad-leaved tree in Northeast China, used cigarette butts to ignite the fuelbed of Q. mongolian leaves with different packing ratios and moisture contents and found that moisture content and wind speed significantly affect ignition probability [11]. Zhang used cigarette butts to ignite the fuelbed of red pine needles and found that, in addition to the influence of meteorological factors and fuelbed characteristics, the position in which the cigarette butts fell also had a significant impact on ignition probability [18]. These studies revealed that the factors that affect the ignition probability of cigarette butts mainly include the moisture content and packing ratio of the fuelbed, wind speed, and falling position. However, these studies did not conduct a systematic analysis, had a small number of factor gradient settings, and had a narrow range, which could not reliably reveal the impact of the factors on cigarette butt ignition.
Research has shown that different types of fuel have distinct physical and chemical properties; therefore, their mechanisms and probabilities of cigarette butt ignition are different. For example, the combustibility of Quercus glauca is not as good as that of P. massoniana, mainly due to the difference in effective mineral content between Quercus glauca and P. massoniana, which are 8.9% and 1.82%, respectively, with a nearly 5-fold difference [19]. Currently, data on the cigarette butt ignition of typical forest fuels are lacking. As the main coniferous tree species in Guizhou Province, P. massoniana has an area of 1.7 million hectares, accounting for about 17% of the province’s forest land, and is rich in oil and highly flammable [2,20]. After the cigarette butts fall, they often ignite the surface layer of dead needles [21,22]. Therefore, the needle layer of P. massoniana was considered the research object in this study, and cigarette butts were used to ignite fuelbeds with different packing ratios and moisture contents under different wind speeds. The factors influencing the ignition probability and time of cigarette butt were determined, and a prediction model for ignition probability and time was established. The study findings provide the informational basis for the study of the ignition mechanisms of other fire sources and promote the study of human-caused fires, which is of great significance for scientific forest fire management.

2. Materials and Method

2.1. Field Investigation and Sample Collection

The weathering times of needles of P. massoniana after withering are different; their physical and chemical properties and their responses to cigarette butt ignition are also different [23]. The fire prevention period in Guizhou Province is from October 1 of a year to May 31 of the following year, with forest fires being most severe from February to April [24]. To ensure representation and practical significance, a P. massoniana forest with a large proportion of forest age (representative) and a south-facing slope (with more dry fuels) was selected as the standard plot. In March 2022, a 20 × 20 m standard plot [25] was set up on Baozhu Mountain, Guiyang City, Guizhou Province, and then measured the characteristics of the forest. Thirty is the minimum sample size with statistical significance, so thirty 10 × 10 cm2 plots were randomly set in the standard plot to measure the average thickness and packing ratio of the P. massoniana fuelbeds to provide a practical basis for indoor bed construction. The basic information on the standard plot and fuelbed is presented in Table 1.
After field investigation, the P. massoniana needles were collected. The different physical properties of needles, such as their surface area-to-volume ratio, may affect the ignition results [26]. Therefore, to avoid the uncertainty of the needle’s characteristics affecting the results, this study tried to ensure that the collected needles were structurally intact and consistent and brought them back to the laboratory for the burning experiment.

2.2. Preparation of P. massoniana Needle Bed Indoors

Experimental ignition of P. massoniana needles with cigarette butts requires prior preparation of needle bed layers with different moisture contents and packing ratios. The specific preparation method was as follows.
(1) Fuelbed of P. massoniana with different packing ratios. A topless iron frame with an area of 15 cm2 was selected to support the fuel. According to the field investigation (Section 2.1), the average thickness of a pine needle bed in the field was 10.00 cm. To ensure that the constructed fuelbed best represented the actual situation in the field, the bed thickness was set to 10 cm; thus, the volume of the fuelbed was equal to the bed area multiplied by the thickness, which was 15 cm2 × 10 cm = 150 cm3 = 0.00015 m3. The range of the packing ratio of the fuelbed in the field was 0.016–0.067; therefore, in this study, five fuelbed packing ratios were selected: 0.016, 0.027, 0.038, 0.049, and 0.067.
The packing ratio indicates the compactness of the fuelbed, which is defined as the fraction of the fuel array volume that is occupied by fuel [26]. The greater the value is, the more compact the needles are in the needle bed [27]. The packing ratio of the bed was mainly determined by the volume density of the fuelbed and particle density of the P. massoniana needles [26], as shown in Equation (1). According to Equation (1), the mass of fuel required for different packing ratios was obtained, and the fuelbed layer of P. massoniana with the corresponding mass was weighed.
β = ρ a ρ b = m v ρ b ,
where, β denotes the packing ratio, ρ a denotes the volume density of fuelbed (kg·m−3), ρ b denotes the particle density of P. massoniana needles (constant value, 536.52 kg·m−3 [2]), v denotes the fuelbed volume (0.00015 m3), and m denotes the quantity of P. massoniana needles (kg).
(2) Fuelbed of P. massoniana with different moisture contents. Through preliminary experiments and the reference literature, it was found that when the moisture content of P. massoniana needles exceeded 15%, ignition of the needles was difficult [11]. Therefore, in this study, four levels of moisture content were used: 0%, 5%, 10%, and 15%. Prepare the bed of needles with different moisture contents for the ignition experiment. Place the P. massoniana needles in an oven and dry them until their quality does not change, recording the dry weight. According to the formula for calculating moisture content, the ratio of the wet weight minus the dry weight to the dry weight and the quality of water required to be added to the needle with different moisture content can be obtained (e.g., if the dry weight is 100 g and the moisture content is 5%, then 100 × 15% = 5 g of water needs to be added). Place the dried needles on the floor, quickly spray the required quality of water, and then place them in a sealed bag for 24 h to allow the needles to fully absorb the water. Prepare a fuelbed using needles with the moisture content required for the ignition experiment. Before each ignition experiment, a subsample of fuel was taken, and its moisture content was tested using a moisture rapid analyzer (ML50, A&D Company, Tokyo, Japan) to ensure an error of less than 1% [11,28].

2.3. Ignition Experiment Using Cigarette Butts

The wind velocity in forests generally does not exceed 6 m/s [29], and when the wind velocity is greater than 5 m/s, cigarette butts generally cannot stay on the surface of the fuelbed [11]; therefore, the ignition experiments were conducted at six different wind velocities: 0, 1, 2, 3, 4, and 5 m/s. An electric fan was used to create the airflow, a handheld weather station was used to measure the wind velocity at the center of the bed, and the distance between the fan and fuelbed was adjusted to obtain different wind velocities.
The most common type of cigarette in the study area was selected: Guiyang Hard Gaozun, a flue-cured tobacco with a nicotine content of 1 mg and a circumference of 25 mm. A burning cigarette butt (including a filter tip length of 2.5 cm, red state) was dropped from a 1.3 m high cylindrical tube onto the corresponding wind velocity position of the fuelbed, ensuring the repeatability of each burning experiment. Based on the classification conditions of fuelbed packing ratio, moisture content, and wind velocity, there were a total of 5 (packing ratio) × 4 (moisture content) × 6 (wind velocity) = 120 sets of conditions, and 30 repetitions (30 is the minimum sample size with statistical significance) were conducted for each combination; therefore, a total of 3600 ignition experiments were conducted. A schematic diagram of the experiment is shown in Figure 1. Fuelbed ignition was considered successful after the cigarette butt fell onto the fuelbed, the needles were ignited, and the fire continued to spread; this was recorded as 1, and the other outcomes were recorded as 0. The indoor air temperature and humidity were recorded for each ignition experiment, and a mobile phone was used to shoot the entire experimental process.

2.4. Data Processing

2.4.1. Analysis of Ambient Air Temperature and Relative Humidity

The average, minimum, and maximum values of the indoor air temperature and relative humidity during the entire ignition experiment were analyzed (Table 2), and it was determined whether the entire ignition experiment was conducted in an environment with relatively constant air temperature and humidity using the coefficient of variation (CV). When the CV is less than 0.3, it indicates that the variability of air temperature and relative humidity is small and can be considered to be in a constant environment [30,31].
The coefficients of variation for air temperature and relative humidity were 0.20 and 0.14, respectively, suggesting that the entire ignition experiment was conducted in a relatively constant environment.

2.4.2. Statistical Ignition Probability

The number of successful ignitions under different levels for each influential factor was calculated to determine the ignition probability at that level. For instance, the fuelbed packing ratio had five levels, and 3600/5 = 720 cigarette butt-ignition experiments were conducted for each level. The number of successful ignitions under each level was recorded as “n”, regardless of the wind speed and water content. “n” divided by 720 was denoted as the ignition probability of that level. A line graph with the influential factor as the horizontal coordinate and the ignition probability as the vertical coordinate was drawn, and the influence of a single factor on the ignition probability was derived.

2.4.3. Impact Factor Analysis

This section mainly analyzes the influence of the following factors (packing ratio, wind velocity, and moisture content) on ignition probability and time. To avoid the impact of redundant data on factor analysis, firstly, the dataset was manually reviewed line by line, and no data segments with high similarity were found, indicating no redundancy issues.
Factors influencing ignition probability were analyzed as follows. This study had five packing ratio levels, four moisture content levels, and six wind velocity levels. Each level of each factor is combined one by one, resulting in a total of 120 sets of combinations, each of which underwent 30 repeated experiments; therefore, each combination had an ignition probability (number of ignitions/30). With ignition probability as the dependent variable and bed-packing ratio, moisture content, and wind velocity as independent variables, a variance analysis was conducted to identify the factors that had a significant impact on the ignition probability of cigarette butts.
Factors influencing the ignition time were analyzed as follows. From the data of successful ignition, the average ignition time for each combination was calculated, with the average ignition time as the dependent variable and bed packing, moisture content, and wind velocity as the independent variables. A variance analysis was performed to determine the factors that had a significant impact on ignition time.
Thereafter, using the significant, influential factors as the horizontal axis and the ignition probability or time as the vertical axis, we analyzed how these factors affected the two important indicators.

2.4.4. Prediction Model of Ignition Probability

The logistic regression method (LRM), general linear method (GLM), and nonlinear regression method (NLM; essentially, it is a self-built method that selects the appropriate model form based on the influence of factors on the ignition probability) were used to establish a prediction model for the ignition probability of cigarette butts. All methods randomly used 70% of the data as a training set to establish the model, and the remaining 30% were used as a test set for validation. For the LRM, the dataset for modeling and validation comprises 3600 ignition data (the independent variable is the factor, and the dependent variable is 0 for no ignition and 1 for successful ignition). For the GLM and NLM, the datasets for modeling and validation are 120 sets of combination data (the independent variable is the factor, and the dependent variable is the ignition probability). The specific methods used were as follows.
(1) LRM. Logistic regression is widely used in forest fire probability models, as shown in Equation (2), to evaluate the binary classification problem of occurrence and non-occurrence [32,33]. In this study, ignition was considered to have occurred, whereas non-ignition was considered to not have occurred. Using the fuelbed packing ratio, moisture content, and wind velocity of the training set data as independent variables and ignition as the dependent variable, the stepwise regression of the “glm” function in R (version 2023.04) was used for logistic modeling, and the test set was used for verification. A receiver operating characteristic (ROC) curve was drawn, and the area under the curve (AUC) was calculated by loading the “pROC” package. When the AUC exceeds 0.7, the prediction accuracy requirements are generally believed to have been met [34,35].
P = e b 0 + b 1 β + b 2 m + b 3 w 1 + e b 0 + b 1 β + b 2 m + b 3 w ,
where P denotes the ignition probability (%), e denotes the natural logarithm, b 0 – b 3 denote the model coefficients, β denotes the fuelbed packing ratio, m denotes the moisture content of the fuelbed (%), and w denotes the wind velocity (m/s).
(2) GLM. Using ignition probability (Section 2.4.2) as the dependent variable and fuelbed packing ratio, moisture content, and wind velocity as independent variables, randomly selecting 70% of the data as the training set data, using Minitab 2021 to selecte the general linear regression method to establish a prediction model, and using the remaining 30% of the data as the test set data for validation, the mean absolute error (MAE) and root mean square error (RMSE) of the model were calculated using Equations (3)–(5) to evaluate the prediction effect of the model.
M A E = 1 n ∑ i = 1 n X i − X j ,
R M S E = ∑ i = 1 n ( X i − X j ) 2 n ,
where X i denotes the predicted value, X j denotes the measured value, and n denotes the test-set data.
(3) According to the situation of the influential factors in Section 2.4.2, moisture content and wind velocity have a significant impact on the ignition probability. First, using the moisture content as the independent variable, based on the influence of moisture content on the ignition probability at each wind velocity level, different equation forms ( P = a m + b , P = a e x p ( m − b ) ,   P = a ( m − b ) 2 , etc., where P indicates the ignition probability (%), m indicates the moisture content (%), a and b indicates the model parameters) were selected to attempt to establish a prediction model for the probability of ignition at each wind velocity. The model with the lowest MAE value was considered to be the appropriate model, and the prediction model for the ignition probability at each wind velocity level was obtained as P = a m + b . Taking wind velocity as the independent variable, the relationship between the parameters a and b of the previous model and wind velocity is analyzed. The model form of a and b based on wind velocity are both selected as a / b = c ( w − d ) 2 + e (where a or b indicates the parameters of the previous model; w indicates the wind velocity; and c, d, and e indicate the model parameters). Finally, the parameter prediction model based on wind velocity was brought into P = a m + b , and a prediction model for the ignition probability based on moisture content and wind velocity was obtained. The MAE and RMSE values of the model were calculated to evaluate the model’s performance.
Draw a 1:1 graph with the measured value as the horizontal coordinate and the predicted value as the vertical coordinate. If there is no error in prediction, the measured and the predicted values are exactly the same, each scatter falls exactly on the 1:1 line, and the regression line perfectly coincides with the 1:1 line. The greater the model error, the farther the scatter deviates from the 1:1 line and the greater the angle between the regression line and the 1:1 line. This study analyzed the prediction effect of the three types of ignition probability prediction models in different ignition probability intervals by plotting a 1:1 graph. The measured values of the LRM were 0 and 1, so the LRM selected ignition probabilities under different combinations as the measured value, and the values calculated based on LRM were used as the predicted values.

2.4.5. Prediction Model of Ignition Time

The average ignition time for each combination was calculated, with the average ignition time as the dependent variable and the fuelbed packing ratio, moisture content, and wind velocity as independent variables. A GLM was selected to establish a prediction model for the ignition time of cigarette butts. We selected 70% of the data as the training set to establish the model and the remaining 30% as the test set for validation. The MAE, RMSE, and mean relative error (MRE) of the model were calculated using Equations (3)–(5), respectively.
M R E = 1 n ∑ i = 1 n X i − X j X j × 100 %
where X i denotes the predicted value, X j denotes the measured value, and n denotes the test-set data.

3. Results

3.1. Basic Situation Statistics of Ignition Probability

Figure 2 shows the variations in the ignition probability of cigarette butts with changes in the fuelbed moisture content, packing ratio, and wind velocity. The ignition probability first increased, then decreased, and then increased with changes in the packing ratio, showing no evident rule. Furthermore, as the wind velocity increased, the probability of ignition first increased and then decreased, and the moisture content of the fuelbed layer had a retarding effect on the ignition probability.

3.2. Analysis of Ignition Probability

3.2.1. Analysis of Influencing Factors

Table 3 shows the results of the variance analysis of ignition probability. The fuelbed packing ratio had no significant effect on the ignition probability, whereas wind velocity and moisture content had very significant effects on the ignition probability.
The packing ratio had no significant effect on ignition probability. Therefore, in this study, the packing ratio under different moisture content and wind velocities was regarded as a repeated treatment, and the influence of different wind velocities and moisture content on the ignition probability was analyzed. When the wind velocity was 0 m/s, regardless of the moisture content, the cigarette butt could not ignite the fuelbed. When the wind velocity was 1–2 m/s, the ignition probability generally decreased with the increasing moisture content; however, no significant differences were noted between different moisture content levels. With an increase in wind velocity, the ignition probability decreased significantly with an increase in moisture content. When the moisture content was 15%, regardless of the wind velocity, the cigarette butt could not ignite the P. massoniana fuelbed. When the moisture content was <15%, the probability of cigarette butt ignition first increased and then decreased with an increase in wind velocity; moreover, with an increase in moisture content, the effect of wind velocity on the probability of ignition became less significant (Figure 3).

3.2.2. Prediction Model of Ignition Probability

The LRM, NLM, and GLM were selected to establish a prediction model of ignition probability by cigarette butts for the P. massoniana fuelbed (Table 4). The NLM had the smallest prediction error of only 0.587%, whereas the GLM had a larger error of 1.880%. The error of the LRM was 1.702%, and the AUC of the test set was 0.757 (Figure 4).
Figure 5 shows a 1:1 plot of the measured and predicted values of the three ignition probability prediction methods. The NLM had the best prediction effect, followed by the LRM and GLM. When the ignition probability was <2.5%, the predicted values were generally high, whereas when the ignition probability exceeded 2.5%, the predicted values were low.

3.3. Analysis of Ignition Time

Table 5 shows the basic information on the ignition time throughout the entire experimental process. It can be seen that the ignition time varies with a range of 164–435 s, the standard deviation is 57.64 s, and the CV is 0.214.

3.3.1. Analysis of Influencing Factor

The time required to ignite the fuelbed after coming in contact with the cigarette butt was significantly affected by the fuelbed packing ratio, wind velocity, and moisture content (Table 6).
As the packing ratio increased, the ignition time showed a decreasing trend, which became evident when the ignition time of the sparsely dense pine-needle bed was significantly higher than that of other density levels, and the difference in ignition times was not significant when the packing ratios were higher. The ignition time first decreased and then increased with an increase in wind velocity. When the wind velocity was 4 m/s, the ignition time was the shortest and was significantly higher than that at lower wind velocities. As the moisture content of the fuelbed increased, the ignition time increased significantly (Figure 6).

3.3.2. Prediction Model of Ignition Time

The prediction model for the ignition time of P. massoniana fuelbed was t = 359.4 − 916 β − 18.95w + 3.60m (R2 = 0.362), with MAE, RMSE, and MRE of 38.09s, 48.85 s, and 14.04%, respectively. Figure 7 shows the 1:1 plot of the prediction of ignition time. When the ignition time was short, the predicted value was higher, and when the ignition time was long, the predicted value was lower.

4. Discussion

4.1. Number of Ignition

In this study, 3600 cigarette butt-ignition experiments, in which 85 needle beds of P. massoniana were ignited, with an ignition probability of 2.36%, were conducted. Zhang studied litter in Mongolian Oak and Pinus koraiensis forests in Northeast China and used cigarette butts to ignite them 2160 times. Mongolian Oak broadleaves and P. koraiensis needles were ignited 72 and 39 times, respectively, with corresponding ignition probabilities of 3.33% and 1.81% [18]. The fuelbed of P. massoniana is more difficult to ignite than the broadleaf bed of Mongolian Oak because of the differing leaf shapes of the broadleaves and conifers [36,37]. Through repeated cigarette butt-throwing experiments, it was found that cigarette butts tended to first smolder, gradually carbonize the litter, and finally transform that smolder into an open flame with the action of wind [38]. Compared with the broadleaves of Mongolian oak, the P. massoniana needles have a smaller area of contact and a smaller initial stage of smoldering, which is more difficult to ignite and produce a flame [26].
However, the ignition probability of P. massoniana needles is higher than that of P. koraiensis needles, mainly because the former are more prone to combustion. The physical and chemical properties of the flammable needles of P. massoniana and P. koraiensis were characterized via laboratory tests and literature investigations, as shown in Table 7 [3]. The calorific value of P. massoniana needles was higher than that of P. koraiensis needles, and the effective and total mineral contents of P. massoniana needles were lower than those of P. koraiensis needles [26,39].

4.2. Influence Factor

The moisture content of P. massoniana fuelbed had a significant effect on ignition probability and time. When the moisture content exceeded 10%, regardless of the changes in other conditions, the cigarette butts could not ignite the fuelbed. Through ignition experiments, Sun et al. found that the maximum ignitable moisture content of broadleaves was 15% [11], and that of coniferous leaves was 10% [29]. Our study findings are similar, and it can be considered that cigarette butts could only ignite coniferous leaves when the moisture content of the coniferous bed was below 10%. As the moisture content increased, the ignition probability decreased; however, no significant differences between adjacent moisture content levels were noted. The time for the cigarette butts to ignite the fuelbed showed a significant increasing trend, with an increase in moisture content, and a significant difference was observed between the different moisture content levels. This may be because the cigarette butt, as a heat source, must evaporate the moisture of a needle before ignition. The lower the moisture content of the needle, the shorter the ignition time required and the easier it is to ignite the needle, resulting in a higher ignition probability [40].
Wind velocity had an extremely significant effect on ignition probability and time. In the no-wind condition, P. massoniana needles could not be ignited regardless of the moisture content and packing ratio of the fuelbed, which is similar to the conclusions drawn by previous studies [29,41,42]. This is mainly because the cigarette butt smolders, which also make the fuel smolder after contact. Smoldering itself cannot produce a flame, and only under the action of certain airflows can it be converted into a flame [43]. Zhang ignited needles of P. koraiensis and found that ignition was possible only when the wind velocity was >2 m/s [18], whereas the minimum wind velocity that resulted in ignition in this study was 1 m/s. As previously mentioned, this was mainly due to the fact that the needles of P. massoniana were easier to ignite. With an increase in wind velocity, the ignition probability of the P. koraiensis fuelbed by cigarette butts first increases and then decreases, and the ignition time first decreases and then increases; moreover, the inflection point occurs when the wind speed is 4 m/s, which is consistent with the results of Yang et al. [38]. However, some scholars believe that with an increase in wind velocity, the ignition probability continues to increase, and the time required for conversion into a flame decreases [5,44]. This is mainly due to the duality of the effect of wind velocity on ignition [11,45]; on the one hand, wind velocity reduces the temperature in the combustion zone and inhibits ignition, and on the other hand, it increases the oxygen concentration in the combustion zone and transforms smoldering into a flame. Regarding the influence of wind velocity, different scholars have reached different conclusions, which may be due to different wind velocity level settings. The ignition probability can be predicted to decrease when the wind velocity gradient is sufficient. Sun et al. used cigarette butts to ignite a broadleaf bed with an optimal wind speed of approximately 3.7 m/s [11], similar to our study findings. Therefore, it can be concluded that when the wind velocity is approximately 4 m/s, cigarette butts are most likely to ignite forest surface fuels and cause a disaster.
The packing ratio of the fuelbed had no significant effect on the ignition probability but had a significant effect on the ignition time. Satoh et al. studied the needle bed under Japanese pine forests and found that as the packing ratio increased, the ignition probability increased by about 50%, whereas Sun et al. reported that the packing ratio of broadleaf beds had no significant effect on ignition probability [11]. These results mainly differed due to the different effects of packing ratios on ignition probability. Similar to wind velocity, the fuelbed packing ratio also has a dual effect on ignition probability; with an increase in the packing ratio, the degree of contact between fuel monomers in the fuelbed is conducive to heat transfer and accumulation; however, with an increase in fuel density, it affects the oxygen concentration in the combustion zone and inhibits combustion [45]. In this study, with an increase in the packing ratio, the ignition probability of the cigarette butts also showed an increasing trend; however, no significant differences were noted between the different packing ratio levels (Figure 2). When the packing ratio continues to increase, the ignition probability can be predicted to exhibit a decreasing trend.

4.3. Prediction Model

Three types of prediction methods were selected to establish the ignition probability model; the fitted line between the measured and predicted values of the NLM is closest to the 1:1 line, and the residual values for different measured values are the closest to zero. The LRM is second, and the GLM has the worst prediction effect. In this study, according to the effects of moisture content and wind velocity on the ignition probability, an appropriate equation form was selected to obtain the ignition probability prediction model (NLM). The prediction error of the model was only 0.587%, which is significantly lower than that of the other methods, and it can reveal the influence mechanism of wind velocity and moisture content on ignition probability, which is more practical and suitable as a cigarette butt ignition probability prediction model. A general linear equation was used to establish a prediction model for the ignition time. The MAE and MRE of the model were 38.09 and 14.04%, respectively, with MRE being below 15%, indicating that the error is within an acceptable range of error [46,47,48]. In future research, we should establish a prediction model with physical significance based on the influence mechanism of the factors and reduce the error and extrapolation ability, which has more practical applications.

4.4. Forest Fire Management

In this study, the ignition probability of the cigarette butt was only 2.36%, and the occurrence of fire required a certain wind velocity and <10% moisture content, which are relatively harsh conditions in the field environment. According to a literature review, the needles of P. massoniana in southwest China are relatively flammable [2], whereas the litter of other typical forests in the study area, such as the Quercus glauca, Quercus acutissima, Quercus fabri, and Phyllostachys edulis forests, are more difficult to burn [3]. However, cigarette butt-induced forest fires account for a vast majority of human-caused fires. For example, in the study area, it is second only to sacrificial fire and children playing with fire [4], indicating the existence of some loopholes in fire source management, especially in outdoor smoking management. Therefore, it is necessary to strengthen fire source management and forest fire prevention publicity to reduce the occurrence of human-caused fires. In this study, the shortest time for the cigarette butt to ignite the fuelbed was 164 s (2.7 min), whereas the longest time was 435 s (7.25 min). Therefore, when the cigarette butt falls on the fuelbed of P. massoniana, a flame appears and spreads in approximately 7 min at most. In forest fire management, extinguishing flames within 7 min is of great significance in preventing forest fires caused by cigarette butts.
In this study, only cigarette butts were selected as the ignition source, and the fuelbed of P. massoniana was selected as the research object for analysis. However, different fire sources and fuels exhibit different ignition probabilities. In future studies, the analysis of the ignition probability and time of other fuel types should be performed using various human-caused fire sources under different conditions to provide data and technical support for the improvement of forest fire prediction models. In addition, as this study was conducted indoors, a knowledge gap remains for the ignition in the actual situation in the field. Moreover, the ignition probability of cigarette buts is also affected by air temperature and humidity. In future research, it will be possible to conduct ignition experiments by controlling different temperature and humidity levels to analyze the impact of temperature and humidity on the ignition probability. Combined with activities such as planned field burning, cigarette butts will be used to ignite needles to verify the applicability and effectiveness of the model.

5. Conclusions

The effects of moisture content, fuelbed packing ratio, and wind velocity on ignition probability and time of fuelbed of P. massoniana were examined for cigarette butt ignition. In this study, the total ignition probability of the cigarette butts of P. massoniana fuelbeds was 2.36% (3600 experiments were conducted, 85 of which were ignited). The fuelbed of P. massoniana could not be easily ignited by cigarette butts and could only be ignited and spread when the moisture content of the fuelbed was lower than 10% and the wind velocity was over 1 m/s. The time required for cigarette butt ignition ranged from 2.70 to 7.25 min. With an increase in the moisture content of the fuelbed, the ignition probability decreased significantly, and the ignition time increased significantly. With an increase in wind velocity, the ignition probability first increased and then decreased, and the ignition time first decreased and then increased. The optimal wind velocity was approximately 4 m/s. The effect of the packing ratio on the ignition probability was not significant; however, a positive correlation was observed, and the ignition time decreased significantly as the packing ratio increased. A model for ignition probability and time prediction was established, and the error was within an acceptable range. Calculating the potential ignition probability and time of cigarette butts in the study area based on known meteorological value and fuel conditions and carrying out target fire source management patrols are of great significance for scientific forest fire management and other fire source research.

Author Contributions

Y.Z. conceived and designed the experiments, analyzed the data, prepared figures and tables, authored and reviewed drafts of the paper, and approved the final draft. L.T. performed the experiments, prepared figures and tables, and approved the final draft. All authors have read and agreed to the published version of the manuscript.

Funding

This research was jointly supported by China National Natural Science Foundation (Grant no.: 32201563), Guizhou Provincial Science and Technology Projects (Grant no.: Qianke Support [2022] General 249), and the intelligent forest fire innovation team of higher education institutions in Guizhou Province (Grant no.: QJJ [2023]075).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy issues.

Acknowledgments

We extend our gratitude to Aixia Luo, Tiantian Sun, and Xuehui Du for their support in indoor work.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

LRM: Logistic regression method, GLM: General linear method, MAE: Mean absolute error, MRE: Mean relative error, NLM: Nonlinear regression method.

References

  1. Wan, F.X.; Zhang, J.C. Ecological Characteristics and Vegetation Rehabilitation Techniques in the Karst Mountain Areas of Guizhou Province. J. Nanjing For. Univ. (Nat. Sci. Ed.) 2003, 27, 45–49. [Google Scholar]
  2. Lan, S.X.; Hu, Z.L.; Jin, Y.Q.; Chen, F.F.; Zhang, Y.L. Effect of Compactness of Fuelbed on Dynamic Change of Moisture Content of Litter Bed of Pinus massoniana. J. Northeast For. Univ. 2022, 50, 45–50. [Google Scholar]
  3. Zhang, Y.L.; Tian, L.L. Examining and Reforming the Rothermel Surface Fire Spread Model under No-Wind and Zero-Slope Conditions for the Karst Ecosystems. Forests 2023, 14, 1088. [Google Scholar] [CrossRef]
  4. Zhang, Y.L.; Guo, Y.; Hu, H.Q. Characteristics of Forest Fire Data in Southwest China during 2001–2017. J. Northwest For. Univ. 2021, 36, 179–186. [Google Scholar]
  5. Manzello, S.L.; Cleary, T.G.; Shields, J.R.; Yang, J.C. On the ignition of fuel beds by firebrands. Fire Mater. 2006, 30, 77–87. [Google Scholar] [CrossRef]
  6. Manzello, S.L.; Cleary, T.G.; Shields, J.R.; Maranghides, A.; Mell, W.; Yang, J.C. Experimental investigation of firebrands: Generation and ignition of fuel beds. Fire Saf. J. 2008, 43, 226–233. [Google Scholar] [CrossRef]
  7. Scholten, R.C.; Jandt, R.; Miller, E.A.; Rogers, B.M.; Veraverbeke, S. Overwintering fires in boreal forests. Nature 2021, 593, 399. [Google Scholar] [CrossRef]
  8. Barforoush, A.S.; du Preez, M. Quantifying the Effectiveness of a Mesh in Mitigating Burning Capabilities of Firebrand Shower. Fire 2022, 5, 150. [Google Scholar] [CrossRef]
  9. Cawson, J.; Pickering, B.; Filkov, A.; Burton, J.; Kilinc, M.; Penman, T. Predicting ignitability from firebrands in mature wet eucalypt forests. For. Ecol. Manag. 2022, 519, 120315. [Google Scholar] [CrossRef]
  10. Fard, M.G.H.; Hostikka, S. Combustion characteristics of non-charring polymer cylinders-experimental and numerical study. Combust. Flame 2022, 249, 112587. [Google Scholar] [CrossRef]
  11. Sun, P.; Zhang, Y.L.; Sun, L.; Hu, H.; Guo, F.; Wang, G.; Zhang, H. Influence of fuel moisture content, packing ratio and wind velocity on the ignition probability of fuel beds composed of Mongolian oak leaves via cigarette butts. Forests 2018, 9, 507. [Google Scholar] [CrossRef]
  12. Zhang, R.Q.; Tong, Y.M.; Wang, X.; Liu, Y.C.; Zhang, Z.Y.; Ruan, Z.C. Study on cigarette butts igniting a Plagiomnium acutum layer in the understory of a Cupressus torulosa forest. J. For. Environ. 2023, 43, 659–665. [Google Scholar]
  13. Hyungsoo, S.; Yeonsook, C. Vulnerability of Pinus dbuttiflora to forest fire based on ignition characteristics. J. Ecol. Field Biol. 2010, 3, 16–20. [Google Scholar]
  14. Kim, D.H.; Lee, M.B.; Viegas, D.X. Ignition of surface fuels by cigarette in forest fire. In Proceedings of the 6th International Conference on Forest Fire Research, Coimbra, Portugal, 15–18 November 2010; Volume 11, pp. 15–18. [Google Scholar]
  15. David, T.; Jeffrey, P.; Douglas, S. Investigation of the decline in reported smoking-caused wildfires in the USA from 2000 to 2011. Int. J. Wildland Fire 2014, 23, 790–798. [Google Scholar]
  16. Markalas, S. Laboratory experiments on the role of burning cigarette-ends as an ignition cause of forest fires. Criminol. Public Policy 2005, 3, 139–160. [Google Scholar]
  17. Satoh, K.; Zhong, Y.L.; Yang, K.T. Study of Forest Fire Initiation Due to Lighted Cigarette—Measurement and Observation of Flaming Probability of Dried Leaves; Report of National Research Institute of Fire & Disaster; National Research Institute of Fire & Disaster: Tokyo, Japan, 2003; pp. 146–153. [Google Scholar]
  18. Zhang, Y.L. Burn Probability of Fuelbed Composed of Mongolian Oak and Pinus Koraiensis Leaf Ignited by Cigarette Ends. Master’s Thesis, Northeast Forestry University, Harbin, China, 2016. [Google Scholar]
  19. Liu, W.L. Prediction of Moisture Contents of Typical Fuels in Nanchang, Jinagxi Province: Optiomation of Models and Evaluation of FWI. Master’s Thesis, Northeast Forestry University, Harbin, China, 2014. [Google Scholar]
  20. Ding, B. Study on the Cause and Regulation of Low-Function Forest of Pinus massoniana, Cunninghamia lanceolata in Guizhou. Ph.D. Dissertation, Guizhou University, Guiyang, China, 2018. [Google Scholar]
  21. Nhongo, E.J.S.; Fontana, D.C.; Guasselli, L.A.; Bremm, C. Probabilistic modeling of wildfire occurrence based on logistic regression, Niassa Reserve, Mozambique. Geomat. Nat. Hazards Risk 2019, 10, 1772–1792. [Google Scholar] [CrossRef]
  22. Bergado, J.R.; Persello, C.; Reinke, K.; Stein, A. Predicting wildfire burns from big geodata using deep learning. Saf. Sci. 2021, 140, 105276. [Google Scholar] [CrossRef]
  23. Anderson, H.E. Moisture diffusivity and response time in fine forest fuels. Can. J. For. Res. 1990, 20, 315–325. [Google Scholar] [CrossRef]
  24. Hu, H.Q. Forest Ecology and Management; China Forestry Press: Beijing, China, 2005. [Google Scholar]
  25. Liu, Y.; Zhang, R.; Lin, C.-F.; Zhang, Z.; Zhang, R.; Shang, K.; Zhao, M.; Huang, J.; Wang, X.; Li, Y.; et al. Remote sensing of subtropical tree diversity: The underappreciated roles of the practical definition of forest canopy and phenological variation. For. Ecosyst. 2023, 10, 100122. [Google Scholar] [CrossRef]
  26. Rothermel, R.C. A Mathematical Model for Predicting Fire Spread in Wildland Fuels; Research Paper INT-115; USDA Forest Service, Intermountain Forest and Range Experiment Station: Ogden, UT, USA, 1972. [Google Scholar]
  27. Nelson, R.M. Reaction times and burning rates for wind tunnel headfires. Int. J. Wildland Fire 2003, 12, 195–211. [Google Scholar] [CrossRef]
  28. de Groot, W.J.; Wardati; Wang, Y. Calibrating the fine fuel moisture code for grass ignition potential in Sumatra, Indonesia. Int. J. Wildland Fire 2005, 14, 161–168. [Google Scholar] [CrossRef]
  29. Jin, S.; Zhang, Y.L.; Zhu, K.; Hu, Y. Burning Probability of Fuelbed Composed of Mongolian Oak Leaf Ignited by Cigarette Butts. J. Northeast For. Univ. 2014, 42, 75–78. [Google Scholar]
  30. Gao, L.H.; Sun, W. Application of variation factor in reliability engineering. J. Acad. Armored Force Eng. 2004, 18, 5–8. [Google Scholar]
  31. Wu, M.; Gu, S.S. The statistical inference of variation coefficient of sample and its applications. J. Tongren Univ. 2010, 12, 139–144. [Google Scholar]
  32. Garcia, C.; Woodard, P.; Titus, S.; Adamowicz, W.; Lee, B. A Logit model for predicting the daily occurrence of human caused forest-fires. Int. J. Wildland Fire 1995, 5, 101–111. [Google Scholar] [CrossRef]
  33. Anderson, P.L.; Loftsgaarden, D.O.; Bradshaw, L.S. Evaluation of fire danger rating indexes using logistic regression and percentile analysis. Int. J. Wildland Fire 2003, 12, 213–222. [Google Scholar]
  34. Padilla, M.; Vega-Garcia, C. On the comparative importance of fire danger rating indices and their integration with spatial and temporal variables for predicting daily human-caused fire occurrences in Spain. Int. J. Wildland Fire 2011, 20, 46–58. [Google Scholar] [CrossRef]
  35. Jiménez-Valverde, A. Insights into the area under the receiver operating characteristic curve (AUC) as a discrimination measure in species distribution modeling. Glob. Ecol. Biogeogr. 2012, 21, 498–507. [Google Scholar] [CrossRef]
  36. Christmann, T.; Rosado, B.H.P.; Delhaye, G.; Matos, I.S.; Drummond, J.S.; Roland, H.L.; Moraes, Y.C.; Menor, I.O. Functional assembly of tropical montane tree islands in the Atlantic Forest is shaped by stress tolerance, bamboo presence, and facilitation. Ecol. Evol. 2021, 11, 10164–10177. [Google Scholar] [CrossRef]
  37. Nocentini, A.; Kominoski, J.S.; O’Brien, J.J.; Redwine, J. Fire intensity and ecosystem oligotrophic status drive relative phosphorus release and retention in freshwater marshes. Ecosphere 2022, 13, 4263. [Google Scholar] [CrossRef]
  38. Yang, G.; Ning, J.; Shu, L.; Zhang, J.; Yu, H.; Di, X. Spotting ignition of larch (Larix gmelinii) fuel bed by different firebrands. J. For. Res. 2022, 33, 171–181. [Google Scholar] [CrossRef]
  39. Garlough, E.C.; Keyes, C.R. Influences of moisture content, mineral content, and bulk density on smoldering combustion of ponderosa pine duff mounds. Int. J. Wildland Fire 2011, 20, 589–596. [Google Scholar] [CrossRef]
  40. Larjavaara, M.; Kuuluvainen, T.; Tanskanen, H.; Venäläinen, A. Variation in forest fire ignition probability in Finland. Silva Fenn. 2004, 38, 253–266. [Google Scholar] [CrossRef]
  41. Ganteaume, A.; Guijarro, M.; Jappiot, M.; Hernando, C.; Lampin-Maillet, C.; Pérez-Gorostiaga, P.; Vega, J.A. Laboratory characterization of firebrands involved in spot fires. Ann. For. Sci. 2011, 68, 531–541. [Google Scholar] [CrossRef]
  42. Ellis, P.F. The likelihood of ignition of dry-eucalypt forest litter by firebrands. Int. J. Wildland Fire 2015, 24, 225–235. [Google Scholar] [CrossRef]
  43. Ganteaume, A.; Lampin-Maillet, C.; Guijarro, M.; Hernando, C.; Jappiot, M.; Fonturbel, T.; Pérez-Gorostiaga, P.; Vega, J.A. Spot fires: Fuel bed flammability and capability of firebrands to ignite fuel bed. Int. J. Wildland Fire 2009, 18, 951–969. [Google Scholar] [CrossRef]
  44. Plucinski, M.P.; Anderson, W.R.; Bradstock, R.A.; Gill, A.M. The initiation of fire spread in shrubland fuels recreated in the laboratory. Int. J. Wildland Fire 2010, 19, 512–520. [Google Scholar] [CrossRef]
  45. Plucinski, M.P.; Anderson, W.R. Laboratory determination of factors influencing successful point ignition in the litter layer of shrubland vegetation. Int. J. Wildland Fire 2008, 17, 628–637. [Google Scholar] [CrossRef]
  46. Zhang, J.Q. Measurement of seasonal changes in moisture content of different tree species. For. Fire Prev. 2000, 2, 17–18. [Google Scholar]
  47. Jin, S.; Yan, X.J. Dynamics and modeling of water content of ten shrub species in their growth period in Maoershan Mountain region of Northeast China. Chin. J. Appl. Ecol. 2012, 23, 3317–3324. [Google Scholar]
  48. Xiao, Y.D.; Zhang, X.Q.; Ji, P. Modeling Forest Fire Occurrences Using Count-Data Mixed Models in Qiannan Autonomous Prefecture of Guizhou Province in China. PLoS ONE 2015, 10, e0120621. [Google Scholar] [CrossRef] [PubMed]
Fire 07 00065 g001
Figure 1. Schematic diagram of the experiment.
Figure 1. Schematic diagram of the experiment.
Fire 07 00065 g001
Fire 07 00065 g002
Figure 2. Variations in ignition probability with changes in the fuelbed moisture content, packing ratio, and wind velocity.
Figure 2. Variations in ignition probability with changes in the fuelbed moisture content, packing ratio, and wind velocity.
Fire 07 00065 g002
Fire 07 00065 g003
Figure 3. Variations in ignition probability with changes in wind velocity and moisture content. Note: Different letters indicate significant differences at the 0.05 level.
Figure 3. Variations in ignition probability with changes in wind velocity and moisture content. Note: Different letters indicate significant differences at the 0.05 level.
Fire 07 00065 g003
Fire 07 00065 g004
Figure 4. ROC curve of the logistic regression method.
Figure 4. ROC curve of the logistic regression method.
Fire 07 00065 g004
Fire 07 00065 g005
Figure 5. A 1:1 plot of the measured and predicted values of ignition probability.
Figure 5. A 1:1 plot of the measured and predicted values of ignition probability.
Fire 07 00065 g005
Fire 07 00065 g006
Figure 6. Variations in ignition time with changes in the fuelbed moisture content, packing ratio, and wind velocity. Note: Different letters indicate significant differences at the 0.05 level.
Figure 6. Variations in ignition time with changes in the fuelbed moisture content, packing ratio, and wind velocity. Note: Different letters indicate significant differences at the 0.05 level.
Fire 07 00065 g006
Fire 07 00065 g007
Figure 7. A 1:1 plot of measured and predicted values of ignition time.
Figure 7. A 1:1 plot of measured and predicted values of ignition time.
Fire 07 00065 g007
Table 1. Basic information about the standard plot and fuelbed of P. massoniana.
Table 1. Basic information about the standard plot and fuelbed of P. massoniana.
Forest
Type		Slope Position Slope
(°)		Mean DBH (cm)Mean Tree Height (m)Canopy DensityAverage Fuelbed
Thickness (cm)		Average Fuelbed Packing RatioForest
Age (year)
P. massonianadownhill13.6020.6014.800.7310.000.03836
Table 2. Basic temperature and humidity conditions.
Table 2. Basic temperature and humidity conditions.
IndexMinimum ValueAverage
Value		Maximum Value25%
Quantile		75%
Quantile		Coefficient of Variation
Air temperature (°C)17.3024.5639.7020.5028.700.20
Relative humidity (%)55.1066.8798.5462.6073.300.14
Table 3. Results of variance analysis of ignition probability.
Table 3. Results of variance analysis of ignition probability.
IndexIII SSdfMSFSig.
Intercept668.7301668.730158.970<0.000
Packing ratio35.17948.7952.0910.087
Wind velocity350.503570.10116.664<0.000
Moisture content 306.0363102.01224.250<0.000
Error450.1101074.207
Total1810.557120
Table 4. Results of prediction model of ignition probability.
Table 4. Results of prediction model of ignition probability.
TypeModelMAE (%)RMSE (%)
Logistics method (LRM) P = e − 3.884 + 0.303 w − 0.150 m 1 + e − 3.884 + 0.303 w − 0.150 m 1.7022.156
General linear regression method (GLM) P = 2.483 + 0.818 w − 0.286 m 1.8802.317
Nonlinear regression method (NLM) P = 0.048m ( w − 3.700 ) 2 − 0.398 m − 1.096 ( w − 3.500 ) 2 + 7.704 0.5870.727
Table 5. Basic information on the ignition time.
Table 5. Basic information on the ignition time.
IndexMinimum ValueAverage ValueMaximum Value 25% Quantile 75 Quantile Standard DeviationCoefficient of Variance
Ignition time (s)164.000269.160435.000230.500300.0057.640.214
Table 6. Results of the variance analysis of ignition time.
Table 6. Results of the variance analysis of ignition time.
IndexIII SSdfMSFSig.
Intercept3,948,959.19413,948,959.19416,921.656<0.000
Packing ratio24,995.04046248.76026.777<0.000
Wind velocity35,364.87748841.21937.885<0.000
Moisture content 8053.46724026.73417.255<0.000
Error7467.75032233.367
Total6,437,273.00085
Table 7. Physicochemical properties of the needles of two conifer tree species.
Table 7. Physicochemical properties of the needles of two conifer tree species.
Tree Species ρ p H (kj·kg−1)ST (%)Se (%)
P. massoniana262.525209,07.2222.4201.820
P. koraiensis316.51217,636.1023.6952.982
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhang, Y.; Tian, L. Ignition of Forest Fires by Cigarette Butts: Using Pinus massoniana Needles as an Example. Fire 2024, 7, 65. https://0-doi-org.brum.beds.ac.uk/10.3390/fire7030065

AMA Style

Zhang Y, Tian L. Ignition of Forest Fires by Cigarette Butts: Using Pinus massoniana Needles as an Example. Fire. 2024; 7(3):65. https://0-doi-org.brum.beds.ac.uk/10.3390/fire7030065

Chicago/Turabian Style

Zhang, Yunlin, and Lingling Tian. 2024. "Ignition of Forest Fires by Cigarette Butts: Using Pinus massoniana Needles as an Example" Fire 7, no. 3: 65. https://0-doi-org.brum.beds.ac.uk/10.3390/fire7030065

Article Metrics

Back to TopTop