An Alternative Method for Dating of Wildfire Occurrence from Tree-Ring Cores: A Case Study in Northeastern Asia

Abstract

The tree-ring fire scar stands as a pivotal proxy for reconstructing historical wildfire occurrences, providing invaluable context for comprehending contemporary wildfire activities during the Anthropocene era. Precise identification of fire scars often necessitates sampling complete tree disks. Yet, stringent forest resource protection policies limit such collection in certain regions, thus curtailing the application of tree-ring-based fire reconstruction. While current methods based on tree-ring cores can ascertain the age range of fire events, pinpointing the exact year remains challenging. In this study, we propose a novel approach for detecting fire occurrence years by recognizing abrupt shifts in the tree-ring geometric center (TRGC). This method entails extracting a minimum of three tree-ring cores from the uninjured side of the tree and in proximity to the fire scar. We validated this method’s efficacy using samples from Transbaikal of Russia, in northeastern Asia, where fire years have already been documented. Our results show that this method accurately identified the fire year in nine fire scars with a 67 percent probability of exact agreement with the actual fire year. It is noteworthy that this method particularly excels in cases of trees with a single fire scar. We recommend employing the traditional method of collecting tree-ring cores near the fire scar to establish the age range of the fire scar. Subsequently, within this determined range, we suggest detecting the shifts in the TRGC to accurately pinpoint the exact year of the fire scar.

1. Introduction

Tree rings are widely used as a proxy for temperature reconstruction of the past millennium, providing a crucial scientific foundation for global warming studies [1]. In addition, tree rings have been extensively utilized for reconstructing hydroclimate [2,3], internal climate variability [4,5], wind speed [6], and fire occurrences [7,8]. These broad applications take advantage of the accurate dating and rich environmental signals found in tree-ring proxies, such as tree-ring width, density, stable isotopes, and wood anatomy features [9,10]. The majority of tree-ring-based studies rely on tree-ring chronology indices with the age-related growth trend removed. However, effectively retaining low-frequency climate signals remains a challenge since it is difficult to distinguish these signals from age-related growth trends. The Basal Area Index (BAI), which measures the annual growth of the tree cross-sectional area, is advantageous in representing low-frequency growth trends [11]. Nevertheless, the BAI often changes smoothly and exhibits weak high-frequency (annual) variability relative to standard tree-ring chronology indices. This limits its application in calibrating with instrumental climate data for climate reconstructions.

A recent study has emphasized the importance of considering asymmetric growth in different parts of a tree trunk when exposed to uneven stress in different directions, such as wind stress [12]. Trees attempt to maintain a circular shape, as circles maximize cross-sectional area for a given perimeter according to Isoperimetric Inequality [13]. As a result, trees tend to shift the geometric center of their ring circles under environmental stress, rather than becoming flattened. Compared with the traditional BAI method, the new BAI method that considers the tree-ring geometric center (TRGC) is capable of preserving both high- and low-frequency variability, which enhances the utility of the BAI for climate reconstruction. The shifts of the TRGC were found to be sensitive to changes in wind speed and have been used to reconstruct wind speed [12].

The primary aim of this study is to assess the efficacy of the TRGC technique in reconstructing the historical occurrences of fires. Existing methods employed for fire reconstruction necessitate the meticulous dating of fire scars, entailing the collection of complete tree disks or cross-sections to locate all fire scars. Unfortunately, this approach can harm venerable, old-growth trees, and it imposes limitations in regions governed by stringent forest resource conservation regulations. A notable example is China, where a multitude of tree-ring-based climate reconstructions have been disseminated [14]. Nonetheless, only a handful of fire reconstructions based on tree rings have been published due to the challenges associated with obtaining permission to sample fire scars, and notably, these investigations have been confined to northeast China [8,15,16]. This scarcity hampers a comprehensive understanding of the intricate interplay between historical fire dynamics and their underlying driving factors across diverse geographic zones. Prior research endeavors have proposed methodologies for fire reconstruction that involve the collection of tree-ring cores [17] or wedges [18] from fire scar sites. The principle is to collect the tree-ring cores or wedges at the fire scar and then cross-date them with tree-ring cores in other directions to determine the specific age of the fire scar. However, collecting wedges can still cause some damage to trees. Since the scar ring typically becomes distinct from the subsequent healing ring, and given that tree-ring cores are prone to fracturing when penetrating the scar ring, the potential for generating gaps or missing rings emerges. Hence, the precise placement of the boring position becomes crucially important at the apex of the fire scar, ensuring the safeguarding of the scar itself while mitigating the risk of ring breakage—an endeavor that proves to be intricate in practical execution. In instances where the borehole placement veers from accuracy, a discrepancy ranging from one to three years might arise between the deduced fire occurrence year and the actual temporal event, as indicated by previous research [17].

In recent years, there has been a growing focus on the anatomical characterization of wood within fire scars. Studies have demonstrated that the physiological structure of trees undergoes changes when they are subjected to fire damage, leading to the development of extremely narrow tree rings near fire scars in the same year [19,20]. Based on this understanding, we put forth the hypothesis that the existence of these remarkably narrow tree rings during fire incidents would result in disparate growth patterns in all directions within the tree, culminating in abrupt shifts in the orientation of TRGC. To empirically scrutinize this proposition, we applied our innovative methodology to samples sourced from the Transbaikal of Russia, a region situated in northeastern Asia, where the precise ages of the fires have been established.

2. Materials and Methods

2.1. Study Region

The study site was positioned in the proximity of Romanovka village (latitude 53.12–53.65° N, longitude 112.80–113.09° E) within Transbaikal, Russia, situated in the expanse of northeastern Asia, covering an area of about 110,000 hectares (refer to Figure 1a) [7]. This locale is marked by a quintessential continental climate, wherein temperature differentials surge beyond 40 °C between summer and winter. The terrain bears the imprint of aridity, characterized by an annual aggregate precipitation not surpassing 400 mm, a consequence of its location enshrouded in the rain shadow cast by the imposing Baikal Mountains [7]. The bulk of the annual precipitation is confined to the months of July and August, accounting for up to 70% of the total rainfall. Notably, this study enclave ranks amongst the foremost trio of regions boasting the highest radiative vigor of boreal forest fires in the northern hemisphere [21,22].

2.2. Tree-Ring Data

We meticulously searched for stumps marked by fire scars on both sides of the road and extracted partial cross-sections from Scots pine and Siberian larch trees displaying indications of fire scars using a chainsaw (Figure 1b). Subsequently, these samples underwent sanding within the laboratory setting, followed by scanning utilizing a high-resolution Epson scanner (Figure 1c). The measurement of tree-ring widths was carried out with the assistance of image analysis software CooRecorder 9.3 [23]. To effectively discern the TRGC and to observe the influence of fire on tree-ring width, we obtained and analyzed three sequences of tree-ring widths across varying directions, recording their corresponding angles. Among these directions, one was in close proximity to the fire scar, while the other two directions were positioned as distantly as feasible from the fire scar. Subsequent to this, we assigned calendar years using the cross-dating methodology, verifying alignments between extremely narrow and wide rings [24]. The quality control of cross-dating was conducted using the COFECHA program [25], while the generation of a standard chronology was facilitated through the utilization of the ARSTAN program [26]. The precise year of the fire was subsequently ascertained by relying upon the dating of the fire scars [7]. Regrettably, most of the cross-sections were characterized by narrow dimensions, posing challenges when endeavoring to conduct measurements along three distinct directions. In order to mitigate potential errors, we meticulously selected five relatively expansive cross-sections featuring pith for this study, thereby enabling the measurement of tree-ring width across various directions. Among these cross-sections, three exhibited a single fire scar, while the remaining two contained three fire scars.

2.3. Detection of the TRGC

The geometric center of a tree-ring circle can be determined by taking at least three tree-ring cores at known cross-angles between them. In application, Fang et al. proposed a method to determine annual shifts in TRGC via the ordinary least squares method [12]. Notably, they observed that the accuracy of the fitting improved proportionately with an increased number of tree ring core samples [12]. We hypothesize that there are N cores intersecting at the pith O, and the angle of the first core is 0°. The pith of the tree is the origin of the coordinate system (Figure 2). We can define θ_i,j (i, j = 1, …, N) as the cross-angle between the i-th and j-th cores, and rj(t) denote the distance between O and the tree ring of a given year t on the j-th core (Figure 2).

For a given year t, the radius of the circumcircle Rt is expressed as follows:

R_t² = (x_j(t) − A_t)² + (y_j(t) − B_t)²

(1)

where (A_t, B_t) is the center of the circumcircle, and x_j(t) and y_j(t) are the coordinates of the tree-ring of the given year t on j-th core in a Cartesian coordinate system. The expansion of Equation (1) can be expressed as:

x_j(t)² + y_j(t)² = 2A_t x_j (t) + 2B_t y_j (t) + R_t² − A_t ² − B_t ²

(2)

In Equation (2), R, A, B are unknowns; so, when there are more than 3 tree-ring cores, the unknown parameters can be solved.

When 3 cores are bored, Equation (2) can be expressed as:

$$\left\{\begin{array}{c}{x}_1^2+y_1^2=2A{x}_1+2B{y}_1+{ R}^2-{ A}^2-B^2 \\ x_2^2+{ y}_2^2=2A{x}_2+2B{y}_2+R^2-{ A}^2-{ B}^2 \\ x_3^2+{ y}_3^2=2A{x}_3+2B{y}_3+R^2-{ A}^2-{ B}^2\end{array}\right.$$

(3)

The specific solution process is as follows. We use linear equation b = XC to represent Equation (3),

$$\mathrm{where} b=\left\{\begin{array}{c}{x}_1^2+y_1^2\\ x_2^2+y_2^2\\ x_3^2+y_3^2\end{array}\right., X=\left[\begin{array}{ccc}{x}_1& y_1& 1\\ x_2& y_2& 1\\ x_3& y_3& 1\end{array}\right], C=\left[\begin{array}{c}2A\\ 2B\\ R^2-A^2-B^2\end{array}\right]$$

However, the coordinates (x, y) are not directly measured; so, we convert the coordinates to polar coordinates using Euler’s formula:

e^iθ = cosθ + isinθ

(4)

In Formula (4), multiplying both sides by r_j(t).

r_j(t) e^iθ = r_j(t) cosθ_i,j + r_j(t) isinθ_i,j

(5)

As relationship $\left\{\begin{array}{c}{x}_j\left(t\right)=r_j\left(t\right)cos{\theta }_{i,j}\\ y_j\left(t\right)=r_j\left(t\right)sin{\theta }_{i,j}\end{array}\right.$, the formula (5) is

r_j(t) e^iθ = x_j(t) + iy_j(t)

(6)

Take the square of the absolute value of both sides

|r_j(t) e^iθ|² = |x_j(t) + iy_j(t)|² = x_j(t)² + y_j(t)²

(7)

In the above linear equation b = XC, b is calculated by |r_j(t) e^iθ|² according to formula (7), and X is calculated by obtaining real and imaginary parts of r_j(t) e^iθ according to formula (6). So, when we measure r_j(t) and θ for more than three tree-ring cores, we can solve C. That is, we obtain the center (A_t, B_t) and radius (R_t) of the circumcircle of each year.

The program for quantifying the TRGC was coded using the MATLAB platform. We revised the program by adding more notes to the program originally presented in a previous study [12], which is shown in the appendix. The program mainly detected the annual shifts of distance and the angles of the TRGC. The radius of the tree-ring circle was also calculated, which can be used to calculate the basal area increment with consideration of the shifts of the TRGC.

3. Results

The establishment of the standard tree-ring chronology encompassed the temporal range, spanning from 1800 to 2016. Within this interval, a collective count of nine fire events was documented across the chosen five cross-sections (Figure 3). Notably, a heightened frequency of fire occurrences manifested during phases marked by diminished tree-ring width, which was particularly evident in the year 2003 (Figure 3). Such a trend could likely be attributed to the sensitivity of the tree-ring chronology to drought, wherein arid and parched climatic conditions are more conducive to the initiation and propagation of fires.

As shown in Figure 4, our methods can reasonably capture the shift of the TRGC in individual trees. Prior to the occurrence of fire, the TRGC exhibited a consistent trajectory in a specific direction. However, the abrupt alteration in the shift course of the TRGC can be ascribed to the significant external impact of the fire event. In order to investigate the connection between the shift distance of the TRGC and fire events, we computed the year-to-year fluctuations of both the abscissa and ordinate of the TRGC. As shown in Figure 5, in single fire scar trees, the fire event induces not only an abrupt alteration in the migration trajectory of the TRGCs but also an extended shift distance (Figure 5a,c,e). Significantly, the fire year pinpointed using these two concurrent characteristics harmonizes with the actual fire occurrence year (

Table 1. Comparison of fire-scar year estimates from tree-ring cores with those based on cross-sections.

No. of Fire Scars	From Cross-Sections	From Tree-Ring Cores
1	1863	1863
1	1872	1872
1	1922	1922
3	1878, 1903, 1917	1880, 1907, 1917
3	1968, 1987, 2003	1971, 1987, 2003

). This congruence could be ascribed to the pronounced impact of fire on tree-ring width, where the disparity in tree-ring width adjacent to the fire scar and in other directions is noteworthy during the fire year. Among the three cross-sections with single fire scars, two displayed remarkably narrow growth rings in the core proximate to the fire scar (Figure 5b,d), while another exhibited exceptionally broad growth rings (Figure 5f).

4. Discussion

4.1. Abrupt Shifts in TRGC as an Indicator of a Fire Occurrence

Prior investigations have demonstrated that when trees are exposed to windy environments or situated on slopes and subject to external forces such as continuous wind stress or the influence of gravity, uneven growth can transpire within various segments of the trunk [27]. This phenomenon results in the broadening of tree rings on the leeward side or downhill of coniferous trees, and in response, trees tend to readjust their geometric center to ensure vertical growth devoid of deformities [28,29]. Correspondingly, in instances of wildfire exposure, heightened temperatures frequently lead to the demise of the vascular cambium, reducing the vascular lumen diameter and cambium cell counts [20]. Consequently, sections of trees impacted by fire commonly display exceedingly narrow growth rings during the fire year and subsequent years, while segments unscathed in other directions often manifest wider growth rings. Hence, the inception of a fire elicits notable oscillations in tree-ring width, provoking trees to instigate an asymmetric growth pattern (Figure 5b,d), eventually culminating in an abrupt shift of TRGC (Figure 5a,c).

In a minority of cases, the fire may not inflict substantial harm upon the tree but might result in the demise of nearby competitors. During such instances, the tree can absorb a greater quantity of nutrients, potentially leading to accelerated healing near the fire scar, resulting in the development of relatively wide growth rings (Figure 5f) [17]. This scenario could also give rise to asymmetrical tree growth and prompt abrupt transitions in TRGC orientation (Figure 5e). It is noteworthy that a predominant proportion of the fires in the study region transpired during the spring [7], a period when trees were actively producing earlywood. Consequently, fires during this timeframe had a more pronounced impact on tree ring width for that particular year. Conversely, if the majority of fires in the study area had occurred late in the growing season when trees were forming latewood, their influence on tree ring width would have been more significant in the subsequent year [19]. Hence, when employing this method, the interrelation between TRGC azimuth changes and fire occurrences should be analyzed in accordance with the prevailing fire season in the study area.

In trees marked with multiple fire scars, it is imperative to avoid all areas bearing fire scars during the drilling process in order to procure an intact, fully traversing tree-ring core. This precaution arises from the fact that the fire-scarred ring tends to be segregated from the subsequent healing ring, and penetrating a fire-scarred tree while drilling can lead to a fragmented core. The tree-ring core extracted from the pith can thus become compromised, rendering it arduous to traverse proximate to all fire scars along the same trajectory. Nevertheless, alterations in tree-ring width prompted by fire tend to manifest within localized zones [20]. This renders the detection of atypical shifts in tree ring width adjacent to fire scars challenging (Figure 5h,j), consequently causing the sharp reorientation of the TRGC during fire years to be less pronounced (Figure 5g,i). Hence, it is often possible to detect only individual fire events closest to the boring path, making it difficult to identify all fire events.

4.2. Boring Method in the Field

It is essential to highlight that the experiments detailed above were executed using samples from tree cross-sections. Yet, when embarking on fieldwork to collect tree-ring cores, several factors warrant careful consideration. To commence, a thorough assessment of the stand’s composition should be undertaken to identify trees bearing well-defined and intact fire scars, ensuring the wood’s structural integrity for sampling purposes [18]. Once target trees are identified, the collection of tree-ring cores from the undamaged side of the tree (Figure 6, cores 4 and 5) should follow, with scrupulous attention devoted to the direction of boring and meticulous recording of measurements.

When boring a tree-ring core at the fire scar site, it is worth noting that the scar ring often displays alterations in color and width, and the tree-ring core might be prone to breakage at the scar [18]. If the acquired tree ring core exhibits the afore mentioned characteristics, it indicates that the fire scar has not been circumvented (Figure 6, core 1, core 2). At this juncture, the count of tree rings on both sides of the fire scar should be tallied to establish a rough age range for the fire scar. Subsequently, continuous core sampling should be performed in the vicinity of the fire scar until a complete core is obtained (Figure 6, core 3). This process requires close observation for any sudden shifts in tree-ring width within the fire scar’s age range, including the presence of an exceedingly narrow ring. If the procured tree-ring core lacks the previously mentioned attributes, this suggests that the fire scar has indeed been avoided. In such scenarios, the endeavor to bore tree-ring cores near the fire scar must persist. Notably, ensuring the precision of the boring location is paramount, as it directly influences the ultimate outcome. Therefore, a substantial number of tree-ring cores should be bored (Figure 6, core 1, core 2) to accurately pinpoint the position for core 3.

The intricate interrelation between wildfires and environmental dynamics has undergone extensive scrutiny over the past few decades, driven by the escalating scale and severity of wildfires witnessed across numerous global regions. The heightened acknowledgment of rising temperatures, evolving patterns of precipitation, premature initiation of spring snowmelt, and prolonged periods of heightened fire vulnerability accentuate the potentially serious consequences at hand [30,31,32]. Research into the historical records of tree-ring fire occurrences have revealed the teleconnection influences of large-scale ocean-atmospheric circulations on fires. However, such investigations remain notably sparse in certain regions, particularly China, necessitating comprehensive and systematic research efforts [7,8]. Our proposed methodology holds considerable promise in discerning the fire years associated with trees marked by a single fire scar. We advocate the traditional practice of extracting cores proximate to fire scars to ascertain the temporal scope of these scars, subsequently analyzing the TRGC shifts within this timeframe to precisely determine the fire mark’s chronological position (as illustrated in Table 1). While the precision of this approach in pinpointing the exact year of a fire is constrained when dealing with trees with multiple fire marks, a combination of methodologies such as fire mark drilling or wedge techniques can be employed to reliably ascertain the occurrence of all fire events.

5. Conclusions

The primary objective of this study was to evaluate the effectiveness of a methodology focused on abrupt shifts in the TRGC for identifying fire incidents. This method involved extracting a minimum of three tree-ring cores from the undamaged side of the tree, adjacent to the fire scar, while avoiding the complete removal of the tree disk. Our findings demonstrated that for tree disks with a rounded shape and a limited number of fire scars, the TRGC’s orientation indeed underwent an abrupt alteration during the fire year. This methodology shows substantial potential in accurately identifying the fire years associated with trees bearing a single fire scar. However, when trees had multiple fire scars, the method’s ability to detect fire events was comparatively constrained. In such scenarios, we recommend resorting to conventional practices, such as extracting tree-ring cores and wedges near the fire scar, to determine the temporal extent of the fire scar and scrutinize the changes in the TRGC within that timeframe, thereby precisely establishing the year of the fire occurrence. In forthcoming endeavors, it is paramount to develop more efficient strategies for detecting fire events that minimize damage to trees.