Structure and Regeneration Differentiation of Coniferous Stand Groups in Representative Altay Montane Forests: Demographic Evidence from Dominant Boreal Conifers

Abstract

With the intensification of global climate change and human activities, coniferous species as the main components of natural forests in the Altay Mountains are facing the challenges of aging and regeneration. This study systematically analyzed structural heterogeneity and regeneration of three coniferous stand groups, Larix sibirica Ledeb. stand group, Abies sibirica Ledeb.-Picea obovata Ledeb.-Larix sibirica mixed stand group, and Picea obovata stand group, respectively, across western, central, and eastern forest areas of the Altay Mountains in Northwest China based on field surveys in 2023. Methodologically, we integrated Kruskal–Wallis/Dunn’s post hoc tests, nonlinear power-law modeling (diameter at breast height (DBH)–age relationships, validated via R², root mean square error (RMSE), and F-tests), static life tables (age class mortality and survival curves), and dynamic indices. Key findings revealed structural divergence: the L. sibirica stand group exhibited dominance of large-diameter trees (>30 cm DBH) with sparse seedlings/saplings and limited regeneration; the mixed stand group was dominated by small DBH individuals (<10 cm), showing young age structures and vigorous regeneration; while the P. obovata stand group displayed uniform DBH/height distributions and slow regeneration capacity. Radial growth rates differed significantly—highest in the mixed stand group (average of 0.315 cm/a), intermediate in the P. obovata stand group (0.216 cm/a), and lowest in the L. sibirica stand group (0.180 cm/a). Age–density trends varied among stand groups: unimodal in the L. sibirica and P. obovata stand groups while declining in the mixed stand group. All stand groups followed a Deevey-II survival curve (constant mortality across ages). The mixed stand group showed the highest growth potential but maximum disturbance risk, the L. sibirica stand group exhibited complex variation with lowest risk probability, while the P. obovata stand group had weaker adaptive capacity. These results underscore the need for differentiated management: promoting L. sibirica regeneration via gap-based interventions, enhancing disturbance resistance in the mixed stand group through structural diversification, and prioritizing P. obovata conservation to maintain ecosystem stability. This multi-method framework bridges stand-scale heterogeneity with demographic mechanisms, offering actionable insights for climate-resilient forestry.

1. Introduction

Forest ecosystems serve as biodiversity hotspots and plays pivotal roles in climate regulation, soil and water conservation, etc. [1,2]. However, global warming is profoundly reshaping the structure and function of forest ecosystems [3]. The IPCC [4] report highlights that rising global temperatures have increased the frequency of extreme climate events, significantly affecting the tree growth, biodiversity, and carbon storage capacity of forests. Coniferous forests, as typical vegetation in cold regions, are particularly sensitive to changes in temperature and precipitation [5,6,7,8]. Devi et al. [9,10] pointed out that continuous climate change could impact plant growth and disrupt the functions of ecosystems. Similarly, Linder et al. [11] found that warming can affect phenology, alter photosynthesis, and consequently influence plant growth, species composition, forest distribution, and even the overall stability of forest ecosystems. In cold semi-arid regions, forests are now under considerable stress due to their inherent vulnerability and sensitivity to environmental changes. Given these climate-driven pressures, understanding stand structural dynamics—particularly size, age, and regeneration patterns—becomes critical to assessing forest structure and guiding adaptive management.

Stand groups constitute structural entities where conspecific cohorts coexist through niche-mediated associations, forming the fundamental demographic units for population persistence and community succession. Stand structure attributes involve characteristics such as tree size, diameter, height, diversity, etc. [12,13,14,15,16], while variations reveal the temporal and spatial changes in the composition and abundance of stand groups [17,18,19,20]. Diameter at breast height (DBH) distribution serves as a critical indicator of forest stand structure, with variations across DBH classes revealing insights into regeneration status and ecological stability [21]. Research has highlighted the importance of trees with large and extra-large DBH in carbon storage and habitat provision within forest ecosystems [22]. Conversely, a higher proportion of trees with small DBH often signifies active regeneration capacity, with forests undergoing recovery or expansion [23]. Tree height distribution further reflects species competition and forest stratification [24]. Forests composed of mature trees are characterized by complex tree height distributions with multi-layered vertical structures, enhancing resource efficiency and biodiversity [25]. In contrast, forests composed of seedlings and saplings exhibit concentrated lower height distributions, reflecting rapid growth potential and intensified resource competition [26]. The ecological succession theory of Odum [27] underscores the linkage between forest stability and tree age composition. A higher proportion of young stands typically indicates strong regeneration capacity, while a dominance of mature and old stands reflects stability and ecological maturity [22].

Static life table and dynamic index analysis provide further quantitative tools for understanding changes in the survival and mortality of species [23]. Previous studies on species structure and dynamics have commonly utilized DBH class data to construct static life table and develop tools such as quantitative dynamic indices and survival curves [16,28,29,30]. These methods allow for the analysis of current survival status and environmental adaptability while shedding light on natural regeneration patterns and responses to disturbances. However, although DBH data are easy to obtain and operate, they cannot accurately reflect tree age information, which may lead to misunderstandings of the structure and variations in stand groups. Using tree age data by tree-ring analysis via tree core sampling can provide a more precise basis for examining the structure and variations in forest stand groups [13].

The Altay Mountains in Northwest China is an important forest ecosystem, with an average elevation of 2140 m. However, with the intensification of global climate change and human activities, the natural forests in the Altay Mountains of Northwest China face severe ecological pressures, including warming temperatures, shifts in precipitation patterns, etc. Larix sibirica Ledeb., Picea obovata Ledeb., and Abies sibirica Ledeb. are the dominant coniferous species in the Altay Mountains in Northwest China. A. sibirica is an evergreen species that prefers sunlight and is adapted to cool and wet climatic conditions. It is mainly distributed between 1400 and 2400 m in the Altay Mountains of Northwest China. P. obovata is also an evergreen tree species that prefers sunlight and is adapted to cool climate, with an elevation range of about 1200–2500 m. In contrast, L. sibirica is a typical positive pioneer species in the natural distribution area (1300–2600 m in elevation), growing rapidly and exhibiting strong drought resistance. These dominant coniferous species coexist through niche differentiation, reflecting certain similarities (such as adaptation to cool climate) while also exhibiting ecological distribution and functional differences. Research on coniferous species in the Altay Mountains of Northwest China has primarily focused on ecological adaptation, biomass estimation, climatic responses, and anthropogenic disturbance impacts, lacking in-depth analyses of stand structure and regeneration capacity, particularly regarding static life tables and dynamic quantitative analyses, resulting in limitations in assessing population stability.

This study aims to analyze the structure and regeneration of coniferous stand groups in the representative montane forests of the Altay Mountains in Northwest China through survey data from sampling plots. The purpose of this study aims to address the following two questions: (1) How do structural heterogeneity and regeneration capacity differ among dominant coniferous stand groups in the representative montane forests, and what do these differences reveal about their future ecological management? (2) Does the species’ complementarity effect in the mixed stand group (A. sibirica-P. obovata-L. sibirica) enhance the radial growth rate and disturbance resistance? The findings are of significant guiding importance for formulating effective forest management and conservation plans, ensuring the sustainable development of the forest ecosystem of the Altay Mountains.

2. Materials and Methods

2.1. Study Area

The Altay Mountains in Northwest China with a forest area of 1,371,737 ha are located in the northernmost part of Xinjiang, China (Figure 1). The region belongs to the temperate semi-arid continental climatic zone: the western forest area has an average annual temperature of 4.83 °C and an average annual precipitation of 147.67 mm; the central forest area has an average annual temperature of 4.55 °C and an average annual precipitation of 202.65 mm; and the eastern forest area has an average annual temperature of 3.45 °C and an average annual precipitation of 196.17 mm (data from meteorological stations). In recent years, there has been a trend of warming and humidification in the Altay Mountains of Northwest China (Figure A1).

2.2. Field Investigation and Laboratory Analysis

This study analyzed data at the plot scale, with plots distributed across different altitudes in the western, central, and eastern parts of the representative montane forests in the Altay Mountains in Northwest China. Based on the distribution of coniferous forests and dominant coniferous species in the representative montane forests, this study established nine sampling plots (north slope) at three altitude gradients (low, medium, and high) in the western, central, and eastern forest areas for plot surveys, each with an area of 100 m × 100 m. Within each sampling plot, four 25 m × 25 m subplots were laid out, resulting in a total of 36 subplots. In the western forest area (48°22′ N, 87°10′ E), the three sampling plots consist of pure L. sibirica species, and the three altitude gradients are 1500–1600 m (slope of 0–15°), 1900–2000 m (10–25°), and 2200–2300 m (0–15°). In the central forest area (48°00′ N, 88°19′ E), the three sampling plots are composed of A. sibirica, P. obovata, and L. sibirica, and the three altitude gradients are 1600–1700 m (30–40°), 1800–1900 m (10–30°), and 2000–2100 m (10–30°). In the eastern forest area (47°11′ N, 89°59′ E), the three sampling plots consist of pure P. obovata species, and the three altitude gradients are 1200–1300 m (0–15°), 1700–1800 m (0–5°), and 2100–2200 m (0–10°). The detailed locations of the sampling plots are shown in Figure 1. To control the effects of elevation, the elevation gradients in the western, central, and eastern forest areas partially overlap, and the sampling plots were selected to avoid areas with recent human disturbances (e.g., logging, afforestation), ensuring that the assessment reflects natural stand dynamics. Furthermore, to ensure the representativeness of the research design, a random placement method was employed during the plot selection, randomly selecting a representative area to minimize selection bias as much as possible.

Field surveys were conducted from June to August 2023. A GPS device (Zhuolin Technology, Hefei, China; precision: 1 m) was used to record detailed environmental information, including the latitude, longitude, and elevation. In the surveyed subplots, the height and basal diameter of all seedlings and saplings were measured using a tape measure and a vernier caliper (Guilin Guanglu Measuring Instrument Co., Ltd., Guilin, China; precision: 0.02 mm), respectively, and their quantities were recorded. For mature trees, the species and number of all trees were investigated. Tree height was measured individually using a laser rangefinder (Shenweida SW-800; Shenweida Technology (Guangdong) Co., Ltd., Dongguan, China; precision: ±2 mm), and trunk circumference was measured individually with a tape measure at a height of 1.3 m above the ground. During the surveys, fallen trees, tree stumps, and dead seedlings within the subplots were also recorded [31]. In this study, 225 L. sibirica individuals were recorded in the sampling plots of the western forest area. In the sampling plots of the central forest area, 926 coniferous trees were surveyed, including 688 A. sibirica individuals, 110 L. sibirica individuals, and 128 P. obovata individuals. In the sampling plots of the eastern forest area, 913 P. obovata individuals were identified.

Following dendrochronological standards [32], we extracted 1–2 cores per tree at 1.3 m height using a Haglöf increment borer (Haglöf Sweden AB, Långsele, Sweden). In each plot, >20 healthy trees across representative DBH classes were sampled, yielding 80, 139, and 90 cores from the western, central, and eastern forest areas, respectively. The cores were air-dried for seven days, mounted in wooden grooves, and sanded (400- to 800-grit) until ring boundaries were microscopically visible. Ring widths were then measured using LINTAB 6 (Rinntech, Heidelberg, Germany; ±0.01 mm) and cross-dated via COFECHA [33], excluding poorly correlated samples.

2.3. Data Processing and Analysis

2.3.1. Age Structure and Radial Growth

Descriptive statistics of DBH data calculated from field-measured trunk circumference and tree age dating data from tree core samples were determined to understand the relationship between them. According to the biological significance of allometry, the nonlinear quantitative relationship between the size of biological individuals and other attributes is often represented in the form of power function [34,35]. The DBH–age relationships were modeled using power functions (Equation (1)) at the forest area scale for each dominant species. During the modeling process, logarithmic transformation was applied to the data to improve the model’s fit. The model’s accuracy was tested using the following parameters: R², adjusted R², F, root mean square error (RMSE), Bias, and p-value [13].

The power function can be expressed as follows:

$$Y=\alpha D^{\beta }$$

(1)

where Y represents the age of the tree (years); D represents the DBH (cm); and α and β are the coefficients.

The test parameters can be calculated as follows:

$$R^2=1-{\displaystyle \frac{{\displaystyle \sum {\left(y_i-{\overline{\overline{y}}}_i\right)}^2}}{{\displaystyle \sum {\left(y_i-\overline{y}\right)}^2}}}$$

(2)

$$\mathrm{Adjusted} R^2=1-{\displaystyle \frac{\left(1-R^2\right)\left(n-1\right)}{n-p-1}}$$

(3)

$$\mathrm{RMSE}=\sqrt{{\displaystyle \frac{1}{n}}{{\displaystyle \sum _{i=1}^n\left(y_i-\overline{\overline{y_i}}\right)}}^2}$$

(4)

$$\mathrm{Bias}={\displaystyle \frac{1}{n}}{\displaystyle \sum _{i=1}^n\left(\overline{\overline{y_i}}-y_i\right)}$$

(5)

where y_i represents the observed value of the i-th data point; ${\overline{\overline{y}}}_i$ represents the predicted value of the i-th data point from the regression model; $\overline{y}$ represents the mean value of all observed values; n represents the samples; and p represents the number of independent variables in the model.

The DBH–age regression models and test results for different coniferous species in various forest areas are presented in Appendix A,

Table A1. Statistics for regression model between diameter at breast height (DBH) and tree ages of L. sibirica in western forest area.

Statistic	L. sibirica
Mean DBH (cm)	45.3
Maximum DBH (cm)	80.4
Mean age (years)	190
Maximum age (years)	472
n	51
Regression model	Y = 0.1651D1.8496
R2	0.6550
Adjusted R2	0.6466
p	<0.0001
F	77.8479
RMSE (years)	66.0075
Bias	0.7064

Note: The data in the table are the mean ± standard deviation of the values used to establish the regression model. For the model, Y represents the age of the tree and D represents the DBH. RMSE, root mean square error.

,

Table A2. Statistics for regression model between DBH and tree ages of different coniferous species in central forest area.

Statistic	A. sibirica	P. obovata	L. sibirica
Mean DBH (cm)	31.6	45.0	47.3
Maximum DBH (cm)	48.8	58.9	78.8
Mean age (years)	53	70	85
Maximum age (years)	90	112	168
n	30	24	34
Regression model	Y = 1.4813D1.0451	Y = 3.6137D0.7812	Y = 1.162D1.1233
R2	0.7047	0.5625	0.7559
Adjusted R2	0.6913	0.5228	0.7468
p	<0.0001	0.0031	<0.0001
F	52.4978	14.1447	83.5885
RMSE	9.0329	14.4288	19.9755
Bias	0.0152	–0.0563	0.2711

and

Table A3. Statistics for regression model between DBH and tree ages of P. obovata in eastern forest area.

Statistic	P. obovata
Mean DBH (cm)	39.4
Maximum DBH (cm)	84.9
Mean age (years)	75
Maximum age (years)	223
n	50
Regression model	Y = 4.3137D0.7639
R2	0.6025
Adjusted R2	0.5920
p	<0.0001
F	57.5933
RMSE	14.8882
Bias	0.1162

. Then, the measured DBH data were substituted into the models to estimate the ages of trees that were not sampled for cores in the sampling plots, thus obtaining age data for all coniferous trees. Combined with DBH data, this allowed for further analysis of the radial growth distribution characteristics of coniferous stand groups.

2.3.2. Calculation of Static Life Table

In this study, the static life table was established based on the number of individuals in each age class. Since a static life table does not track the entire life history of species in the same period but instead reflects the age distribution of overlapping generations within a species group at a specific time, a smoothing technique was applied to the number of surviving individuals in age class x (A_x) to reduce data volatility [36,37]. The calculation formulas for the static life table are as follows [38]:

$$l_x=a_x/a_0\times 1000$$

(6)

$$d_x=l_x-l_{x+1}$$

(7)

$$q_x=d_x/l_x$$

(8)

$$L_x=\left(l_x+l_{x+1}\right)/2$$

(9)

$$T_x={\displaystyle \sum L_x}$$

(10)

$$e_x=T_x/l_x$$

(11)

$$K_x=\ln l_x-\ln l_{x+1}$$

(12)

$$S_x=l_{x+1}/l_x$$

(13)

where x is the age class; a_x is the correction value of actual number of surviving individuals in x age class; a₀ is the initial value of a_x; l_x is the standardized number of surviving individuals in x age class; d_x is the standardized number of dead individuals from x to x + 1 age class; q_x is the mortality rate from x to x + 1 age class; L_x is the survived individuals from x to x + 1 age class; T_x is the total number of surviving individuals over x age class; e_x is the average life expectancy; K_x is the disappearance rate; and S_x is the survival rate.

2.3.3. Dynamic Indices

This study employed the quantitative analysis method proposed by Chen [39] to describe the differences in coniferous stand groups.

$$V_x={\displaystyle \frac{A_x-A_{x+1}}{\max \left(A_x,A_{x+1}\right)}}\times 100\%$$

(14)

$$V_{pi}={\displaystyle \frac{1}{{\displaystyle \sum _{x=1}^{K-1}{A}_x}}}{\displaystyle \sum _{x=1}^{K-1}\left(A_x\times V_x\right)}$$

(15)

$${V^{\prime }}_{pi}={\displaystyle \frac{{\displaystyle \sum _{x=1}^{K-1}\left(A_x\times V_x\right)}}{K\times \min \left(A_1,A_2,\dots ,A_K\right){\displaystyle \sum _{x=1}^{K-1}\left(A_x\right)}}}$$

(16)

$$P_{\max }={\displaystyle \frac{1}{K\times \min \left(A_1,A_2,\dots ,A_K\right)}}$$

(17)

where V_x is the dynamic index of quantitative change between two adjacent age classes; V_pi is the dynamic index without external disturbance; V′_pi is the dynamic index with external disturbance; A_x and A_x+1 are the numbers of individuals in the x-th and x + 1-th age classes, respectively; K represents the number of age classes; and P_max is the maximum probability of risk assumed by the assemblage, at which point V_pi has the greatest impact on the stand group. When V_x, V_pi, and V′_pi are positive, negative, and zero, they, respectively, indicate growth, decline, and stability in the number of individuals between two adjacent age classes.

2.3.4. Fitting of Survival Curves, Mortality Curves, and Disappearance Rate Curves

The survival curve was plotted with age class x as the horizontal axis and the logarithmic standardized number of surviving individuals in x age class (lnl_x) as the vertical axis. According to the characteristics of survival curve changes, Deevey Jr. [40] classified the survival curve into three types: Deevey-I type (convex), where the species have a low mortality rate among young individuals, and most individuals can survive to the average physiological lifespan, but nearly all die within a short period upon reaching this lifespan; Deevey-II type (diagonal), where the mortality rate is roughly the same across all age classes; and Deevey-III type (concave), where young individuals have a high mortality rate. This study used exponential (n_x = n₀e^−fx, where n_x is the value of lnl_x, x is the age class, and n₀ and f are constants) and power function (n_x = n₀x^−f) equations to determine the Deevey-II curve and Deevey-III curve, respectively [13]. Additionally, the mortality and disappearance rate curves were also plotted.

2.3.5. Data Analysis

This study utilized Excel 2021 and SPSS 29 for data analysis. Excel 2021 was primarily used for preliminary data analysis. SPSS 29, on the other hand, was employed for statistical analysis and correlation tests. First, the Kolmogorov–Smirnov test revealed that the data did not meet the normal distribution assumption; therefore, this study employed the Kruskal–Wallis test to assess the median differences among multiple independent samples. The Kruskal–Wallis test is a non-parametric test with the null hypothesis (H₀) stating that all groups have equal medians, and the alternative hypothesis (H₁) posits that at least one group has a median significantly different from the others. The significance level for the test was set at α = 0.05. If the result of the Kruskal–Wallis test is significant (p < 0.05), a post hoc test was conducted to determine which specific group differs from each other. This study employed Dunn’s t test analysis with Bonferroni-adjusted p-values to control the error rate of multiple comparisons. As a non-parametric pairwise test, Dunn’s t method effectively pinpoints significant intergroup disparities under a consistent significance threshold (α = 0.05).

In the Python 3.13 environment, further data analysis and visualization were carried out using Jupyter Notebook 6.x. Plotting was conducted via the Matplotlib 3.8.x and Statsmodels 0.14.x libraries.

3. Results

3.1. Structural Characteristics

3.1.1. Distribution of DBH

The median DBH of the P. obovata stand group was at 15.57 cm, significantly higher than the 6.35 cm observed in the A. sibirica-P. obovata-L. sibirica mixed stand group (p = 0.000), yet notably lower than the 31.94 cm of the L. sibirica stand group (p = 0.000). The P. obovata stand group exhibited a narrow DBH distribution range, positioning its growth characteristics between the other two stand groups (Figure 2a). By contrast, the mixed stand group exhibited an even narrower range of DBH. The L. sibirica stand group showed the widest distribution range, reflecting its strong growth advantage and adaptability.

For the L. sibirica stand group, trees with a DBH higher than 30 cm accounted for 60.00% of the total, while the frequency of trees in the 5–10 cm DBH class was only 0.89% (Figure 2b). In contrast, the mixed stand group was primarily distributed in smaller DBH classes (frequency of 45.57% for DBH class < 5 cm, and frequency of 19.22% for DBH class of 5–10 cm), and large-diameter trees (DBH > 30 cm) represented a smaller proportion (14.58%). DBH in the P. obovata stand group was relatively evenly distributed within the <25 cm range, with 36.37% of trees having a DBH < 10 cm. Regarding the DBH distribution patterns, the L. sibirica stand group displayed a unimodal distribution, and the mixed stand group followed a decreasing trend in tree numbers as DBH increases. For the P. obovata stand group, tree numbers first fluctuated and then gradually declined with increasing DBH, resulting in an overall unimodal distribution.

3.1.2. Distribution of Tree Heights

The median tree height of the P. obovata stand group (8.50 m) was significantly greater than that of the mixed stand group (2.55 m; p = 0.000) but significantly lower than that of the L. sibirica stand group (16.90 m; p = 0.000). The L. sibirica stand group exhibited the smallest variability in tree height (range of 30.70 m), while the mixed stand group showed greater variability (range of 37.88 m) (Figure 3a). The P. obovata stand group had few trees exceeding 40.00 m in height, reflecting strong growth potential.

The tree height distribution of the L. sibirica stand group was notably concentrated and followed a unimodal pattern (Figure 3b). Tree heights were primarily within the range of 10.0–25.0 m, with no observations below 2.0 m or above 35.0 m. In contrast, shorter trees occupied a significant proportion in the mixed stand group, with 31.21% being ≤ 0.5 m and 15.87% between 0.5 and 2.0 m, the highest proportions among the three stand groups. The P. obovata stand group, however, had a relatively uniform height distribution across all classes, mainly concentrated in the 2.0–15.0 m range.

3.1.3. Distribution of Tree Ages

The median tree age of the L. sibirica stand group (91 years) was significantly higher than that of the P. obovata stand group (35 years; p = 0.000) and the mixed stand group (10 years; p = 0.000). Regarding the distribution range of tree ages, the mixed stand group exhibited the most concentrated age distribution, the P. obovata stand group had a slightly broader age range with a more uniform distribution, whereas the L. sibirica stand group showed the widest distribution (Figure 4).

Due to the fact that the age range of trees in the western forest area was large (minimum of 5 years and maximum of 589 years) while the age range of trees in the central and eastern forest areas was small (minimum of 1 year and maximum of 188 years), we adopted different age classes for the trees (Figure 5). The age structure of the L. sibirica stand group was characterized by a predominance of middle- to old-aged trees, with the frequency of distribution showing a unimodal pattern as tree age increases (Figure 5a). Trees aged 50–100 years accounted for the highest proportion, reaching 40.00%, followed by those aged 1–50 years (15.11%) and 100–150 years (14.67%). The age distribution of the mixed stand group was characterized by a very high proportion of young individuals (1–10 years), accounting for 49.78% (Figure 5b). The P. obovata stand group was dominated by middle-aged trees (10–60 years), comprising 72.95% (Figure 5b). The frequency of tree numbers in the mixed stand group decreased with increasing tree age, while the P. obovata stand group showed a unimodal distribution.

3.1.4. Distribution of Radial Growth

There were highly significant differences in the median radial growth among the three stand groups (p = 0.000). The mixed stand group had the highest radial growth, with an average of 0.315 cm/a, followed by the P. obovata stand group at 0.216 cm/a, while the L. sibirica stand group exhibited the lowest radial growth (0.180 cm/a). The boxplot revealed clear differences in the median radial growth of the three stand groups, ranked as follows: mixed stand group > P. obovata stand group > L. sibirica stand group (Figure 6). The mixed stand group had the narrowest box range and the fewest outliers, the P. obovata stand group had a moderately sized box range with some outliers in the high growth direction, while the L. sibirica stand group showed the widest box range and the most outliers.

3.1.5. Distribution of Density

For the L. sibirica stand group, the peak density was 355 individuals/ha, occurring at an age of approximately 75 years (Figure 7a). After reaching the peak, the density gradually decreased and stabilized as the age increases. For the mixed stand group, the density showed a clear declining trend with increasing tree age (Figure 7b). When the trees aged 1–10 years, the density reached its peak value at 1844 individuals/ha. After 60 years of age, the density declined more significantly and gradually stabilized. The density of the P. obovata stand group peaked (1178 individuals/ha) at a tree age of approximately 38 years (Figure 7b). It then gradually decreased with increasing age, displaying a unimodal distribution pattern.

3.2. Characteristic of Static Life Table

The number of survival individuals in different age classes of the L. sibirica stand group varied significantly (

Table 1. Static life table of different coniferous stand groups.

Stand Group	x (Years)	ax	lx	dx	qx	Lx	Tx	ex	lnlx	Kx	Sx
L. sibirica	1–50	34	1000	–529	–0.529	1265	5529	5.529	6.908	–0.425	1.529
	50–100	52	1529	147	0.096	1456	4265	2.788	7.333	0.101	0.904
	100–150	47	1382	676	0.489	1044	2809	2.032	7.232	0.672	0.511
	150–200	24	706	235	0.333	588	1765	2.500	6.559	0.405	0.667
	200–250	16	471	118	0.250	412	1176	2.500	6.154	0.288	0.750
	250–300	12	353	176	0.500	265	765	2.167	5.866	0.693	0.500
	300–350	6	176	29	0.167	162	500	2.833	5.173	0.182	0.833
	350–400	5	147	59	0.400	118	338	2.300	4.991	0.511	0.600
	400–450	3	88	29	0.333	74	221	2.500	4.480	0.405	0.667
	450–500	2	59	0	0.000	59	147	2.500	4.075	0.000	1.000
	500–550	2	59	0	0.000	59	88	1.500	4.075	0.000	1.000
	550–600	2	59	-	-	29	29	0.500	4.075	-	-
A. sibirica-P. obovata-L. sibirica	1–10	461	1000	469	0.469	766	1736	1.736	6.908	0.632	0.531
	10–20	245	531	262	0.494	400	971	1.827	6.276	0.681	0.506
	20–40	124	269	76	0.282	231	570	2.121	5.595	0.332	0.718
	40–60	89	193	74	0.382	156	339	1.758	5.263	0.481	0.618
	60–80	55	119	63	0.527	88	183	1.536	4.782	0.749	0.473
	80–100	26	56	24	0.423	44	95	1.692	4.032	0.550	0.577
	100–120	15	33	15	0.467	25	51	1.567	3.482	0.629	0.533
	120–140	8	17	7	0.375	14	26	1.500	2.854	0.470	0.625
	140–160	5	11	7	0.600	8	12	1.100	2.384	0.916	0.400
	160–180	2	4	2	0.500	3	4	1.000	1.468	0.693	0.500
	180–200	1	2	-	-	1	1	0.461	0.774	-	-
P. obovata	1–10	116	1000	–560	–0.560	1280	7379	7.379	6.908	–0.445	1.560
	10–20	181	1560	–353	–0.227	1737	6099	3.909	7.353	–0.204	1.227
	20–40	222	1914	52	0.027	1888	4362	2.279	7.557	0.027	0.973
	40–60	216	1862	802	0.431	1461	2474	1.329	7.529	0.563	0.569
	60–80	123	1060	681	0.642	720	1013	0.955	6.966	1.028	0.358
	80–100	44	379	276	0.727	241	293	0.773	5.938	1.299	0.273
	100–120	12	103	-	-	52	52	0.500	4.639	-	-

Note: L. sibirica, Larix sibirica Ledeb.; A. sibirica, Abies sibirica Ledeb.; P. obovata, Picea obovata Ledeb. x, age class; a_x, correction value of actual number of surviving individuals in x age class; l_x, standardized number of surviving individuals in x age class; d_x, standardized number of dead individuals from x to x + 1 age class; q_x, mortality rate from x to x + 1 age class; L_x, survived individuals from x to x + 1 age class; T_x, total number of surviving individuals over x age class; e_x, average life expectancy; K_x, disappearance rate; S_x, survival rate.

). In the first three age classes (1–50, 50–100, and 100–150 years), the number of surviving individuals accounted for 64.89% of the total individuals, indicating the stability of the stand group. In the age classes of 100–150 and 150–200 years, the standardized number of dead individuals reached the highest values, while the mortality rate peaked in the age class of 250–300 years (0.500). The life expectancy fluctuated with increasing age, with trees in the age class of 1–50 years showing the highest life expectancy.

The number of surviving individuals (461) of the mixed stand group was highest in the age class of 1–10 years, and the surviving individuals in the first two age classes (1–10 and 10–20 years) accounted for 68.48% of the total individuals. However, the standardized number of dead individuals peaked in the age class of 1–10 years. As the age increases, the life expectancy first increased and then decreased, with the age class of 20–40 years exhibiting the highest life expectancy.

The P. obovata stand group had the highest number of surviving individuals in the age class of 20–40 years, followed by the age class of 40–60 years, totally accounting for nearly 50.00%. The number of surviving individuals significantly decreased after 80 years of age, and the individual mortality rate reached its maximum value in the age class of 80–100 years. As the age increased, the life expectancy generally showed a declining trend.

3.3. Variation Analysis Between Age Groups

3.3.1. Quantity Variations

For the L. sibirica stand group, V_1–50 and V_150–200 were less than 0.00, V_50–100–V_100–150 and V_200–250–V_400–450 were greater than 0.00, and V_450–500–V_500–550 were equal to 0.00, indicating a “decline-growth-decline-growth-stable” trend within its age class range (

Table 2. Dynamic indices of L. sibirica stand group.

Dynamic Index Class	Dynamic Index Value (%)	Dynamic Index Class	Dynamic Index Value (%)
V1–50	–62.22	V350–400	25.00
V50–100	63.33	V400–450	33.33
V100–150	45.45	V450–500	0.00
V150–200	–18.18	V500–550	0.00
V200–250	63.64	Vpi	39.27
V250–300	12.50	V′pi	1.64
V300–350	42.86	Pmax	4.17

Note: V_1–50–V_500–550, dynamic index of quantitative change between adjacent age classes, with age class shown in Table 1; V_pi, dynamic index without external disturbance; V′_pi, dynamic index with external disturbance; P_max, the maximum probability of risk assumed by the species.

). For the mixed stand group, V_1–10–V_120–140 and V_160–180 were positive, while V_140–160 equaled 0.00, indicating a “growth-stable-growth” trend within its age class range (

Table 3. Dynamic indices of A. sibirica-P. obovata-L. sibirica mixed stand group.

Dynamic Index Class	Dynamic Index Value (%)	Dynamic Index Class	Dynamic Index Value (%)
V1–10	67.46	V120–140	80.00
V10–20	18.00	V140–160	0.00
V20–40	19.51	V160–180	50.00
V40–60	55.56	Vpi	49.73
V60–80	50.00	V′pi	4.52
V80–100	45.45	Pmax	9.09
V100–120	16.67

Note: V_1–10–V_160–180, dynamic index of quantitative change between adjacent age classes, with age class shown in Table 1.

). For the P. obovata stand group, V_1–10 and V_10–20 were less than 0.00, while V_20–40–V_80–100 were greater than 0.00, indicating a “decline-growth” trend within its age class range (

Table 4. Dynamic indices of P. obovata stand group.

V1–10	V10–20	V20–40	V40–60	V60–80	V80–100	Vpi	V′pi	Pmax
–8.66	–57.53	19.73	55.00	81.48	85.00	23.39	1.11	4.76

Note: V_1–10–V_80–100, dynamic index of quantitative change between adjacent age classes, with age class shown in Table 1.

).

Both V_pi and V′_pi were positive for the three stand groups, signifying that the stand groups were all growth-oriented (Table 2, Table 3 and Table 4). Ignoring or considering external disturbance, the growth trend was most pronounced in the mixed stand group, while it was weakest in the P. obovata stand group. Additionally, the mixed stand group had the highest risk probability (P_max = 9.09%), whereas the L. sibirica stand group showed the lowest risk probability (P_max = 4.17%).

3.3.2. Fitting of Survival Curves, Mortality Rate Curves, and Disappearance Rate Curves

The standardized number of surviving individuals in the L. sibirica stand group showed an overall declining trend within the age range of 1–600 year, generally presenting an overall “concave-convex-concave-convex-concave” pattern (Figure 8a). The trends of the mortality rate curve and disappearance rate curve were generally consistent, exhibiting a “growth-decline-growth-decline-growth-decline” pattern. The standardized number of surviving individuals of the mixed stand group generally showed a declining trend within the age range of 1–200 years, indicating an overall “concave-convex-concave-convex-concave” pattern (Figure 8b). The trends of the mortality rate curve and disappearance rate curve all exhibited a “growth-decline-growth-decline-growth-decline-growth-decline” pattern. The standardized number of surviving individuals of the P. obovata stand group first increased slowly and then decreased gradually within the age range of 1–120 years, presenting an overall “convex-concave” pattern (Figure 8c). Both the trends of the mortality rate curve and disappearance rate curve showed an increasing trend.

Table 5. Test model of survival curves for different coniferous stand groups.

Stand Group	Equation	R2	F	P	Survival Curve Type
L. sibirica	nx = 8.1542e−0.062x	0.9496	188.7652	0.000	Deevey-II
	nx = 8.5095x−0.268	0.7073	24.1596	0.001
A. sibirica-P. obovata-L. sibirica	nx = 8.4554e−0.010x	0.8914	73.9427	0.000	Deevey-II
	nx = 22.209x−0.457	0.6007	13.5399	0.005
P. obovata	nx = 7.9925e−0.004x	0.6265	8.3864	0.034	Deevey-II
	nx = 8.9584x−0.085	0.2797	1.9412	0.222

Note: n_x, value of lnl_x; x, age class.

showed that the R² and F values of the exponential function model for all three stand groups were higher than those of the power function model, indicating that the exponential function model provides a better fit, and the survival curves conformed to the Deevey-II type.

4. Discussion

A. sibirica and P. obovata are heliophilous, evergreen coniferous tree species, while the L. sibirica is a heliophilous and deciduous coniferous species. These characteristics determine differences in ecological functions, structures, and regeneration capabilities [7]. Both the A. sibirica and P. obovata are shade-tolerant species, often occupying lower canopy positions in the stand groups, and maintain regeneration in understory environments through strong regenerative capacity. L. sibirica exhibits strong drought resistance, often serving as a pioneer species in establishing communities in arid and nutrient-poor environments [8].

The DBH, tree height, age, and density variations in the L. sibirica and P. obovata stand groups all showed a unimodal distribution, indicating a certain stability in the stand structure. Pretzsch [22] noted that the concentration of tree height and the height range of mature trees often reflect the coordinated development of trees in terms of light exposure, spatial competition, and resource acquisition. The distribution patterns of DBH, tree height, and age indicate that the mixed stand group possesses strong regeneration ability, while the regeneration ability of the L. sibirica stand group is weaker [41]. The mutualistic competition between A. sibirica and P. obovata in the mixed stand group may promote regeneration through resource allocation optimization, while the monospecific stand group is limited by ecological niche overlap of a single species (such as light competition) [42]. However, the lack of large DBH trees within the mixed stand group will also impact its stability. From a forest development perspective, the three stand groups represent a complete successional sequence: the L. sibirica stand group corresponds to a mature community, the mixed stand group reflects an early secondary successional stage, and the P. obovata stand group represents an intermediate successional phase. Thus, the observed spatial heterogeneity in stand structures is essentially a spatial manifestation of temporal succession, making age differences an inherent and meaningful component of the natural forest dynamics.

The survival curves of the three stand groups all exhibit Deevey-II type, which is consistent with the findings of Liu et al. [30] regarding Taxus cuspidata in Northeast China, indicating that the seedling mortality is high in the early stage, while the growth in the following stage tends to stabilize. As time progresses, the mortality rate decreases, leading to a more stable structure in the mid-to-late-stage. Hu et al. [43] indicated that the factors affecting the mortality rate of evergreen coniferous species shift from abiotic to biotic as size increases, whereas the mortality rate of deciduous broadleaf species is primarily influenced by biotic factors. The stability of the number of surviving individuals in different life stages of the L. sibirica stand group further demonstrates its lower mortality rate and higher survival rate, which may be key factors allowing this species to maintain sustainable vegetation development in the face of competition and environmental changes. A study conducted by Kharuk et al. [44] in the Sayan Mountain showed that L. sibirica has strong adaptation potential under future climate conditions due to its drought tolerance. The mixed stand group shows high densities of saplings and seedlings, but also high mortality rates, reflecting its sensitivity to environmental stress and inter- and intra-species competition, a phenomenon closely related to density-dependent mortality mechanisms [27,45]. For heliophilous species, neighboring competition is the main driving force behind trees’ mortality throughout their lifespans [43]. Davis and Condit [46] also noted that tree growth and survival are influenced by neighboring plants, including both resource competition and facilitation, highlighting the importance of neighboring plants in shaping structure. As tree age increases, the density of the three stand groups shows a declining trend. In the late stages, the significant decrease in density poses greater mortality risks, especially under conditions of insufficient resource supply [23,47,48].

The standardized number of dead individuals, individual mortality rate, and disappearance rate in the P. obovata stand group exhibited negative values in the early stage. According to Proctor [49], while negative values in the static life table are inconsistent with mathematical assumptions, they can still provide useful ecological information. This phenomenon may be attributed to density effects. The observed maximum density, reaching 2388 individuals/ha, may be linked to positive density dependence effects, such as group dynamics or local seed dispersal aggregation [50,51]. Although an increase in density may lead to increased competition for resources, resulting in a negative density-dependent effect (such as a decrease in individual survival due to resource scarcity), a positive density-dependent effect can promote the survival and reproduction of species under certain conditions. The high density and robust regeneration capacity of young P. obovata individuals likely explain the emergence of negative values. Despite challenges from resource competition and environmental stress, the species sustains itself through high seed germination rates and rapid replenishment. Thus, it can be inferred that the characteristics of P. obovata in the early stage are characterized by a “rapid replenishment—high elimination” model. Zhang et al. [16] conducted a study on the population of Pinus koraiensis in the Changbai Mountain and found that the population shows a growth type, and the high mortality rate of juvenile individuals, limited living space and resource conditions, and the obvious physiological aging of older individuals are the main reasons restricting the growth of P. koraiensis population, as is consistent with the changes observed in the P. obovata stand group in this study.

The mixed stand group has a significant advantage in radial growth, while the monospecific stand group may perform poorly due to lower resource utilization efficiency or weaker ecological adaptability. In the mixed-species stand group, different tree species exhibit differentiated strategies in light, nutrient, and water utilization, effectively utilizing resources at different levels through niche differentiation, reducing resource competition, and optimizing resource allocation [42]. Compared to monospecific forests, the interactions between different tree species in mixed forests lead to higher growth rates and biomass, especially in resource-limited environments [52,53]. According to the study by Loreau and Hector [54], biodiversity positively impacts the productivity of ecosystems, and the complementarity effects between species significantly increase with species richness. Mixed stand groups enhance the ecosystem’s adaptability to environmental changes by increasing species diversity and niche differentiation, thereby promoting the radial growth of species. The mixed stand group in this study exhibits a variation in “growth-stable-growth”, showing the most pronounced overall growth trend regardless of external disturbances. This stability can be attributed to the species diversity of the mixed-species stand group and its more effective resource utilization and niche differentiation abilities [55]. The growth potential of the monospecific stand group (L. sibirica or P. obovata) is weak, particularly under disturbance conditions, which has a more pronounced negative impact on their species quantity [42]. However, the maximum risk probability of random disturbances shows that the mixed stand group has the highest maximum risk probability, indicating that it is more sensitive to random disturbances. This is because the species composition and structure within the mixed stand group are more complex, and in the face of external disturbances, the responses of different species may have conflicting synergistic and competitive effects [54]. The L. sibirica stand group has the lowest maximum risk probability, indicating a strong ability to withstand random disturbances, which can be attributed to its advantages in niche adaptability in unfavorable environmental conditions [7,56].

Although this study employed a systematic sampling design to capture the characteristics of dominant coniferous stand groups in the representative montane forests of the Altay Mountains in Northwest China, the representativeness of the sample data remains subject to certain limitations. Additionally, due to the absence of long-term, repeated measurements, the static life table analysis provides results that approximate the probability of occurrence for trees within specific size classes rather than direct estimates of survival or mortality rates. To enhance the robustness of future research, efforts should focus on expanding the number of sample plots through comprehensive field surveys, supplemented by remote sensing technology for broader spatial coverage and long-term monitoring [57]. Such an approach should encompass a wider range of ecological gradients and environmental conditions to more accurately elucidate the structural and dynamic characteristics of the dominant coniferous species in the Altay Mountains region.

Furthermore, this study prioritizes stand-scale structural heterogeneity and regeneration dynamics as immediate drivers of conservation planning. However, future research could integrate two complementary axes to deepen actionable insights. First, linking stand complexity patterns (e.g., canopy gaps, species mixtures) with individual tree growth responses across climatic gradients may reveal how structural diversity balances ecosystem stability (e.g., carbon sequestration) and regeneration capacity (e.g., post-disturbance recruitment). Second, coupling these ecological assessments with species-specific climatic thresholds (e.g., drought tolerance, thermal optima) could identify priority areas where targeted interventions (e.g., density management, assisted migration) would most effectively buffer against climate-driven regeneration failures. For instance, high-resolution projections of thermal stress zones, overlaid with stand structural maps, might pinpoint vulnerable slopes where assisted migration of drought-tolerant genotypes could prevent recruitment collapse. Such cross-scale syntheses would align restoration targets with physiological and climatic realities, ensuring that management strategies are both ecologically viable and climate-responsive.

5. Conclusions and Recommendations

5.1. Conclusions

This study systematically explored the structure and regeneration characteristics of the dominant coniferous stand groups across representative montane forests (western, central, and eastern forest areas) of the Altay Mountains in Northwest China. The L. sibirica stand group in the western forest was primarily composed of trees with large DBH, indicating a certain level of stability in the ecosystem in the short term but weak natural regeneration capability in the long term. The A. sibirica-P. obovata-L. sibirica mixed stand group in the central forest area was dominated by trees of small DBH, with young structure and high regeneration capacity. The P. obovata stand group in the eastern forest area showed a uniform distribution in DBH, with a normal but slow regeneration capacity. All coniferous stand groups showed a growth-oriented pattern, and the survival curve conformed to the Deevey-II type. It is recommended to conduct research focusing on more sample plots through artificial field surveys combined with remote sensing technology over larger areas to more comprehensively reveal the structure and variation characteristics of the forests in the Altay Mountains region in the future.

5.2. Management Recommendations

Building on the results, we proposed differentiated protection strategies for different forest areas as follows: (1) Western forest area—designating high-altitude areas as core protection zones for old-growth forests, limiting human disturbances (such as grazing), and promoting seed dispersal and natural regeneration by establishing ecological corridors, while maintaining ancient forest resources; conducting artificial replanting trials in low-altitude areas and regularly monitoring survival rates. (2) Central forest area—monitoring the distribution of A. sibirica seedlings and young trees, and establishing long-term fixed sample plots combined with remote sensing technology to track changes in stand structure in mid-altitude regions; near-natural forest management combining natural thinning and artificial intervention: implementing scientific selective thinning and dynamic density regulation measures, optimizing stand density, reducing overcrowding competition, and promoting rapid and healthy tree growth. (3) Eastern forest area—strictly controlling grazing management in low-altitude areas to promote undergrowth regeneration; promoting healthy growth in high-altitude regions: utilizing stratified thinning methods based on DBH distribution characteristics to retain healthy dominant tree species, removing poorly growing or pest-affected individuals, and optimizing stand structure. Moreover, given the mixed stands’ significantly greater radial growth (0.315 vs. 0.180–0.216 cm/a; p < 0.001) and higher dynamic indices (V_pi = 49.73 and V′_pi = 4.52), our results support adopting mixed-species models to enhance productivity and stability in forest management.