Drivers of Forest Dieback and Growth Decline in Mountain Abies fabri Forests (Gongga Mountain, SW China)

Abstract

Mountains are global biodiversity hotspots but face the danger of habitat loss, especially at lower elevations due to climate-warming-induced forest dieback. In the Gongga Mountains (SW China), Abies fabri trees at 2800 m show increased mortality, yet the causes remain unclear. We assessed climatic influences and bark beetle infestations on tree vigor and radial growth, comparing healthy and declining trees at 2800, 3000, and 3600 m elevations. Leaf nitrogen and phosphorus concentrations were measured to evaluate nutrient status. From 1950 to 2019, mean annual temperatures rose at all elevations, while precipitation decreased at low elevations, negatively correlating with temperature. Such warmer, drier conditions impaired low-elevation trees. The decline in A. fabri growth began in the late 1990s to early 2000s, with an earlier and more pronounced onset at lower elevations. A clear lag is evident, as trees at 3000 m and 3600 m showed either delayed or minimal decline during the same period. High-elevation trees experienced more stable climate and better nutrient availability, supporting greater growth and leaf nitrogen in healthy trees. Bark beetle infestations were worst in declining trees at the highest elevation. Our results reveal that A. fabri vigor shifts along elevation gradients reflect interactions between abiotic and biotic stressors, especially aridification.

1. Introduction

Mountain forests are among the most diverse and threatened terrestrial ecosystems, playing a crucial role in understanding the biosphere’s response to climate change [1]. Forests cover around one-third of all land on Earth and breathe life into our world, but it is not just the planet that suffers when they are destroyed. The ongoing impacts of climate change are reshaping the dynamics of these forest ecosystems, leading to novel temperature gradients, altering resource availability, impacting species dynamics, and modifying disturbance regimes [2,3]. However, many cases of forest dieback have been described in mountain forests, characterized by radial growth decline and elevated mortality rates [2]. In such events of dieback and elevated mortality, multiple drivers act simultaneously and make it challenging to disentangle their relative contributions. However, recent research, such as that by Hammond et al. [4], has identified a “hotter-drought fingerprint” as a consistent global climate signal for tree mortality across various locations and biomes, suggesting a quantifiable commonality in climate anomalies linked to die-off events despite other interacting factors.

In high-elevation mixed-conifer forests across China, researchers have reported mortality events occurring with increased magnitude and frequency over recent decades [5,6,7,8]. Most of these dieback and mortality cases have been linked to environmental stressors, particularly droughts [9,10,11] and also to altered disturbance regimes such as more frequent wildfires [12,13]. These dieback episodes have also been observed in the Henduan mountain chain, which is the main forest area of Southwest China, where the Gongga Mountain (GM hereafter) boasts unique and diverse conifer forest ecosystems still understudied.

Tree growth within diverse GM forests is the result of the interplay between biotic and abiotic factors [14]. Factors such as individual characteristics (size, age, genotype, and functional traits) shape the interaction between trees and their environment [15,16,17,18,19]. Extrinsic factors (including neighborhood interactions, climate [precipitation, temperature, vapor pressure deficit (VPD)], acid rain, herbivory, microbial relations, stand composition/structure, and soil nutrition) further influence forest dynamics and tree growth [20,21,22]. VPD is a critical abiotic factor that affects tree physiology, particularly in the context of drought stress. As temperatures rise due to climate change, VPD increases, leading to higher rates of transpiration and water loss from tree canopies [23,24,25]. This heightened evaporative water demand can result in physiological stress, making trees more susceptible to dieback [26,27]. In regions affected by climate warming, like the GM, forests may face challenges from changing climate conditions [28] because the compounded effects of increased VPD and reduced soil moisture can exacerbate water scarcity, further impacting tree health and growth.

Additionally, the role of other stressors like bark beetles, nutrient deficiencies, and/or acid rain cannot be overlooked. Bark beetles often target weakened trees, which may be suffering from drought stress or nutrient deficiencies [29,30]. As climate change alters the distribution and life cycles of bark beetles and influences tree susceptibility to their attacks, infestations may become more frequent and severe, leading to increased mortality rates of infested stands. The interaction between bark beetles and environmental stressors creates a feedback loop that can accelerate forest decline, highlighting the need for integrated management strategies. Furthermore, the nutrient availability in the soil plays a pivotal role in determining tree vigor. Nutrient deficiencies, particularly in nitrogen (N) and phosphorus, can impair the ability of trees to cope with stressors such as drought and pest infestations [31]. Adequate nutrient levels are crucial for maintaining robust growth and resilience against drought stress. In the case of GM forests, understanding the nutrient dynamics is essential for assessing the overall health of Abies fabri (Masters) Craib populations and their capacity to withstand the pressures of climate change. In addition to these factors, the influence of acid rain has emerged as a significant concern for temperate conifer forests, including A. fabri stands [32]. Acid rain can lead to soil degradation and nutrient leaching, adversely affecting tree growth and health. Such pollution type contributes to the depletion of essential nutrients like calcium and magnesium, which are crucial for maintaining tree vigor [33,34]. Consequently, trees subjected to acid rain may experience increased susceptibility to other stressors such as drought and pest infestations, compounding their risk of dieback [35].

Studies have reported that atmospheric deposition of heavy metals like mercury (Hg) and cadmium (Cd) is intensively accumulated in the forests of GM [36]. While specific reports of “acid rain” as defined by pH values are not explicitly detailed for Gongga Mountain within the provided documents, the accumulation of atmospheric deposition and the discussion of water chemistry in forest ecosystems suggest an awareness and potential for such environmental impacts [37]. For instance, a study in Gongga Mountain analyzed the variation characteristics of pH and ion content in precipitation, throughfall, stemflow, swamp, soil, and runoff within the dark coniferous forest, indicating that throughfall and stemflow absorb acid material, which increases pH values [37]. This implies that acidic components are present in the precipitation. Additionally, other studies discuss acid precipitation as a significant environmental problem in South China, with some regions experiencing stronger acid rain [38]. While these specific regional studies are not directly on Gongga Mountain, they highlight the broader context of acid rain in nearby areas of Southwest China.

Recently, several dieback events affecting A. fabri stands have been observed in the GM [39]. While extensive research has explored various biotic and abiotic stress factors influencing tree growth in diverse coniferous forests across China, a critical gap remains in our understanding of the causes of dieback and high tree mortality. Despite the ecological significance of GM, no study has yet delved into the mechanisms leading to A. fabri dieback and mortality in this region, where there may be a contraction of low-elevation conifer forests subjected to increased drought stress [40]. Therefore, it is essential to understand the causes of A. fabri forest decline in GM, as such understanding will provide crucial insights for their effective management.

To this end, we conducted a thorough evaluation of the dieback processes in GM A. fabri forests, focusing on two key questions: (1) Why are some trees decaying while others remain healthy in the same site? (2) Why does the abundance of declining trees decrease with increasing elevation? We hypothesized that (1) the decline in healthy trees at the lowermost elevation is primarily due to increased drought stress resulting from climate warming; (2) A. fabri growth will be differently impacted by temperature, precipitation, and drought stress along the elevation gradient; and (3) soil and leaf nutrient concentrations and the presence of bark beetle infestations critically impact the overall health of A. fabri forests.

To test these hypotheses, we considered three elevations (2800 m a.s.l., 3000 m a.s.l., and 3600 m a.s.l.). At each elevation level, we took cores to assess radial growth trends of A. fabri trees with different vigor, analyzed climate trends, and correlated climate variables (precipitation, temperature, and a drought index) with growth indices from 1950 to 2019. Additionally, we examined other potential stress factors such as bark beetle infestations and analyzed proxies of vigor such as the concentrations of essential nutrients in both soil and leaves. We also determined the timing of disturbances such as thinning through the quantification of growth releases and suppressions. Ultimately, we discuss both climate–growth relationships and growth releases in the context of shifting climate regimes.

2. Materials and Methods

2.1. Study Area and Study Species

The study was carried out in the GM (29°30′ to 30°20′ N, 101°30′ to 102°15′ E, 1100 m to 7556 m a.s.l) on the southeastern fringe of the Tibetan Plateau, SW China [41,42]. In this area, the climate is characterized by a mean annual precipitation of 2000 mm, with most rainfall occurring between June and October [43]. The study area is located between 29°34′53”, 45° N, 102°00′28”, 66° E and 29°33′02”, 09° N, 102°00′44, 98° E (Figure 1a). The general geology of the study area is represented by a volcanogenic–sedimentary complex. Forests at elevations above 2800 m a.s.l. are predominantly composed of conifer and subalpine vegetation types, including A. fabri and Picea asperata M.T.Masters 1906, which are well-adapted to the harsher climatic conditions found at these altitudes. The understory is dominated by sparse scrubs and herbs, with the cover degree being smaller than 30% [44]. The parent materials of soils are mainly residual-slope deposits of granites and glacial tills [45].

2.2. Experimental Design and Field Data

We conducted field surveys and sampling in August 2019. We collected plant and soil samples at three different elevations (2800, 3000, and 3600 m a.s.l.). A total of nine sites/plots (50 × 50 m) were placed at the three elevations. A. fabri trees were sampled at the three elevation levels, and the distance between the plots was no less than 40 m. Tree vitality was systematically assessed using a multi-criteria decision rule that integrates individual and combined indicators of tree health [46]. Tree vitality was systematically assessed using a multi-criteria decision rule that integrates individual and combined indicators of tree health (crown defoliation, crown structure, and stem condition). The criteria used to distinguish between healthy (H) and declining (D) trees included the following:

• Crown Defoliation: We assessed the level of leaf loss or discoloration in the crown. Healthy trees typically exhibited full, green crowns with minimal leaf loss, while declining trees showed significant defoliation, with reduced leaf density and browning or yellowing of foliage. Healthy trees typically exhibited full, green crowns with minimal leaf loss. Declining trees showed significant defoliation, defined as more than 25% cumulative defoliation, reduced leaf density, and browning or yellowing of foliage.
• Crown Structure: In addition to defoliation, the overall structure of the crown was evaluated. Healthy trees generally had a well-developed, symmetrical crown shape, while declining trees often presented irregularities such as thinning or uneven branch distribution.
• Stem Condition: The health of the tree trunk was also assessed through visual inspections for signs of damage or disease, such as bark lesions, cankers, or signs of decay. Healthy trees typically exhibited robust bark and a strong stem, while declining trees often showed signs of decline, including fungal infections or resin pockets due to bark beetle attacks.

For every sampled tree, the following variables were measured: diameter at breast height (DBH, measured at 1.3 m), height, and health status (H: healthy and D: declining). Tree height was measured using a clinometer, while DBH was measured using a measuring tape. Inspections for bark beetle infestation in each assessed A. fabri tree were conducted through a systematic approach that emphasized visual and physical examination techniques [46]. Each tree was meticulously scrutinized for signs of infestation, including small entry holes and resin pockets indicative of beetle activity, the presence of fine reddish-brown frass surrounding the tree, and excess resin flows (an indication of the tree’s defensive response to beetles) [46,47]. A thorough health assessment of the tree crown, marked by discoloration and defoliation, was also conducted to correlate with infestation levels. Quantitative data was collected by calculating the percentage of infested trees within each plot.

2.3. Climate Data

The climate data used in this study (period 1950–2019) were obtained from the 9 km resolution climate data store “ERA5-Land dataset of Copernicus Climate Change Service [48], including monthly total precipitation and mean minimum and maximum temperatures.

2.4. Tree Core Sample Collection, Processing, and Analysis

We collected cores at 1.3 m from H and D A. fabri trees at each elevation using a 5 mm increment borer. At the laboratory, all cores were prepared (air-dried, glued, and progressively polished) following dendrochronological procedures [49]. Subsequently, they were visually cross-dated and scanned with a resolution of 3200 dpi. Ring widths were measured to the nearest 0.001 mm in CDendro and CooRecorder software (version 9.5) on scanned images [50]. The quality of the cross-dating and measurements was checked using the COFECHA program, which calculates cross-correlations between individual series of each core and a master mean series or chronology [51]. To obtain a more accurate representation of overall tree growth, providing a direct biological interpretation as a measure of wood production and better capturing long-term growth trends, ring-width series were subsequently converted to annual basal area increment (BAI) series using the dplR package in R 4.4.2 [52,53].

Site chronologies (H and D trees separately at different elevations) of tree-ring indices were obtained by detrending ring-width series using non-linear smoothing techniques, specifically, the Friedman Super Smoother spline [54] which captures long-term growth trends and variability more accurately than linear models. First-order autocorrelation was subsequently removed using autoregressive modeling in the dplR package in R [53,55]. This procedure generates a standardized, pre-whitened series of dimensionless indices that preserve a common variance and contain high-frequency variability potentially related to climate. Individual series of indices were averaged using a bi-weight robust mean. The adequacy of the sample size for capturing the hypothetical population signal was assessed by calculating the Expressed Population Signal (EPS), with a threshold value of 0.85 used to identify the “critical year” at which each site chronology became reliable enough and well replicated [56].

To reconstruct past disturbances, we analyzed tree releases and suppressions by employing the radial growth averaging method on ring-width series [57]. We assessed growth disturbances for 15-year running windows. A threshold of 50% growth increase or decrease was established to define a significant release or suppression event. For a growth release or suppression to be classified as such, it must persist for a minimum of 5 consecutive years. These analyses were performed using the dplR [55] and TRADER [58] R packages (version 3.6.3, R Development Core Team, 2024).

2.5. Climate–Growth Relationships

Climate–growth relationships were assessed for the period 1950–2019, considering the climate window from the previous January to the current October, that is, the year of tree-ring formation. We quantified the effects of temperature, precipitation, and drought severity (quantified using 12-month values of the Standardized Precipitation-Evapotranspiration Index (SPEI)) [59] on the interannual variability of ring-width indices through bootstrapping correlations and response function analyses [49]. This was performed by correlating the site mean series of ring-width indices with monthly climatic data.

SPEI was calculated using the SPEI R package [60]. Correlations of chronologies with monthly climate variables and 12-month-long SPEI values were calculated using the treeclim R package [61].

2.6. Soil and Leaf Samples: Collection and Analyses

To characterize the soil physicochemical properties, we randomly collected five soil cores (5 cm × 20 cm, A horizon) within each of the three plots at every elevation level. Samples from the same elevation were pooled together, passed through a 2 mm mesh-size sieve to remove roots and debris, air-dried in the shade, and stored in dry paper bags at room temperature for further analyses. The Bray 1 extraction methods described by Bray et al. and Øien and Selmer-Olsen [62,63] were used for the determination of soil available P (AP) and available N (AN), respectively. Further details on AP and AN analyses are provided in Supplementary Materials (Section S1). Soil-dissolved organic C (DOC) was determined following Sarkadi et al. [64]. Soil dissolved carbon (DC) was determined by first adding 2.5 g of field-moist soil, which was shaken with 25 mL of 0.01 M CaCl₂ (1:10 w/v, soil/solution ratio) in 50 mL centrifuge tubes. The tubes were tightly sealed and put on a reciprocating shaker at 200 rpm for 2 h [64]. The soil extracts were then centrifuged at 8000× g for 10 min, and the recovered supernatant and DOC were measured using a TOC/TN analyzer (Tekmar Dohrmann Apollo 9000, Spectralab Scientific Inc., Markham, ON, Canada) using combustion (680 °C) with a platinum catalyst.

To assess leaf nutrient concentrations, we randomly selected 15 H and 15 D trees in each plot. We then sampled 10 fully developed current-year leaves per tree following Cornelissen et al. [65]. At the laboratory, collected current-year leaves were dried at 60 °C for 48 h. The dried samples were then separately weighed, powdered, passed through a 100-mesh screen, and then stored in a cool, dry place. After the acid digestion of samples following Pequerul [66], the leaf concentrations of N and phosphorus (P) were also determined by using acid-base titration and spectrophotometry (SpectraMax iD3 Multi-Mode Microplate Reader, Molecular Devices, Salzburg, Austria), respectively. (Supplementary Materials (Section S1)). Total plant C concentration was analyzed using a CHN analyzer (MT-5, Yanaco Co., Ltd., Kyoto, Japan).

2.7. Data Analyses

The normality of variables was checked using Shapiro–Wilk tests. Brown–Forsythe test statistic ANOVA (F*-test), when sample sizes are equal, and Welch’s ANOVA (W-test), when sample sizes are unequal, were performed with elevation and healthy status as factors to find out elevational changes in several variables (soil and leaf nutrient concentrations; tree DBH and height; health status). The significance of differences between means was verified using Tukey’s HSD (when homogeneity of variance was fulfilled) or Games-Howell post hoc tests (when Levene’s test was statistically significant) [67]. To assess long-term trends in climatic variables, such as precipitation, a cubic polynomial regression model was fitted to the data using the Ordinary Least Squares (OLS) method. All analyses and visualizations were carried out with R Language and Environment (version 4.4.2, R Development Core Team 2024) with base packages, OriginPro 2021 (Origin Lab., Northampton, MA, USA), and Inkscape v.1.0 (Draw Freely-Inkscape).

3. Results

3.1. Climate Trends

Modeling using OLS regression with a cubic polynomial fit, we found that during 1950–2019, mean annual values of precipitation significantly decreased by 9.2 mm at 2800 m (Figure 2a) and showed insignificant reductions of 0.6 mm at 3000 and 3600 m (Figure 2c,e), whereas temperature significantly increased at all elevations, with rises of +0.1 °C at 2800 m (Figure 2b) and +0.2 °C at both 3000 and 3600 m (Figure 2d and Figure 2f, respectively).

Notably, precipitation showed a marked decline in the early 2000s at 2800 m, while temperature exhibited a significant increase, particularly since the 1980s. Additionally, SPEI only significantly (p < 0.001) decreased by 0.02 at 2800 m (Figure 2g). We also found a significant (p < 0.05) increase in VPD was observed, with an overall increase of 0.002 kPa at 2800 m (Figure 2j). This increase was more pronounced at higher altitudes, with VPD significantly (p < 0.0001) increasing by 0.3 kPa at 3000 m (Figure 2h) and 0.4 kPa at 3600 m (p < 0.0001) (Figure 2l).

Seasonal analysis revealed significant mean temperature increases at 2800 m during the summer and autumn (+0.01 °C) and also in winter (+0.02 °C) (Figure S1b–d). At higher elevations, increases were more pronounced in summer and autumn (+0.02 °C), and winter (+0.04 °C) (Figure S1g,h). In the case of precipitation, no significant changes were found (Figure S2). Significant changes in SPEI were also noticed at different elevations (Figure S3). At 2800 m, SPEI significantly increased by 0.02 during spring and winter (Figure S3a,d) and significantly declined by 0.02 during summer and autumn (Figure S3b,c). SPEI only significantly changed in autumn by 0.01 (decline) (Figure S3g). At 3600 m, significant changes in SPEI were also noticed in autumn (0.01 decline) (Figure S3k) and winter (0.009 increase) (Figure S3l). We also noticed significant increases in spring, summer, and autumn VPD at all elevations, and also in winter at 2800 m and 3600 m (Figure S4).

3.2. Changes in Tree Size and Beetle Infestation Along the Elevational Gradient

Tree DBH and height were different across the different elevations, regardless of tree health status (Figure 3). The highest mean DBH and height values corresponded to the mid-elevation site (3000 m), with significant DBH differences observed across elevation levels (Figure 3a). The lowest mean DBH and height values were found at the highest elevations corresponding to D trees. In terms of tree height, significant differences (p < 0.0001) across elevations were also observed (Figure 3b), and tree height also peaked at 3000 m.

Bark beetle infestations were observed at all elevation levels, with significant variations across the elevation gradient (p < 0.0001) (Figure 3c). Most D trees were infested by bark beetles. At the highest elevation, all D trees were infested. In stark contrast, only 9.8% of A. fabri H trees were infested at 3000 m.

3.3. Growth Trends and Disturbances

Moreover, tree-ring chronologies were developed for D and H A. fabri across the three elevations (

Table 1. Tree-ring chronology statistics by health class (D = declining, H = healthy) and elevation (2800–3600 m), showing sample size, timespan, ring width metrics, and quality indicators (EPS, PC1 variance).

Tree Health Class	Elevation (m)	No Trees	Timespan	Mean Tree-Ring Width (mm)	SD (mm)	First-Order Autocorrelation	Mean Sensitivity	Correlation with Master	Variance PC1 (%)	EPS	Best-Replicated Timespan
D	2800	16	1884–2019	2.14	0.91	0.83	0.19	0.36	38.50	0.85	1923–2019
H	2800	15	1892–2019	2.51	0.98	0.77	0.21	0.27	37.10	0.82	1946–2019
D	3000	11	1838–2019	2.20	0.80	0.74	0.22	0.30	55.40	0.98	1929–2019
H	3000	17	1904–2019	2.18	0.71	0.77	0.16	0.37	39.30	0.87	1965–2019
D	3600	16	1856–2019	1.21	0.50	0.82	0.17	0.45	42.30	0.88	1920–2019
H	3600	19	1829–2019	1.19	0.47	0.78	0.17	0.37	39.00	0.86	1924–2019

). The number of sampled trees ranged from 11 to 19, with timespans spanning from the early 19th century to 2019. Mean ring widths varied between 1.19 mm and 2.51 mm, with standard deviations indicating moderate variability across groups. First-order autocorrelation values ranged from 0.74 to 0.83, reflecting varying degrees of year-to-year growth dependency. We also found that D trees were more sensitive to climate. Mean sensitivity, a measure of growth responsiveness, was generally low to moderate (0.16–0.22), slightly higher in declining trees. Correlations with master chronologies were moderate (0.27–0.45), supporting the reliability of the records. Quality statistics showed EPS values above 0.8 for all chronologies, indicating strong signal strength, and PC1 accounted for 37.1% to 55.4% of variance, with higher elevations around 3000 m displaying greater variance explained. Best-replicated timespans spanned mainly the 20th century into recent years, ensuring robust chronology development.

At both 2800 m and 3600 m, the BAI for A. fabri H trees showed insignificant differences in the growth patterns (p > 0.05) (Figure 4a,c), and significant differences in growth pattern were noticed at 3000 m (Figure 4b). Specifically, at the lowest elevation, the growth of A. fabri D trees experienced a decline from the 1950s through the late 1970s, followed by a period of recovery and increase from the 1980s to 2019. In contrast, A. fabri H trees exhibited a consistent upward growth trend throughout the study period. At mid-elevation, the BAI of A. fabri D trees remained higher than that of A. fabri H trees from 1950 until the mid-2010s, during which A. fabri H trees grew more (Figure 4b). Similarly to 2800 m, both H and D trees at 3600 m showed comparable growth patterns. However, A. fabri (H) trees began to exhibit superior BAI from the early 1990s onwards (Figure 4c, and Figures S5 and S6).

At 2800 m, multiple periods of growth suppression, particularly in the 1970s, were recorded in H and D trees (Figure 5a). At 3000 m, A. fabri D trees exhibited suppressions in the 1960s and again in the 2000s (Figure 5b). At 3600 m, suppressions again affected most trees from 1964 to 1978 and again from 1999 to 2004 (Figure 5c).

3.4. Climate–Growth Relationships

We developed coherent chronologies for H and D trees (EPS > 0.85 since 1950) in the three elevation groups (Figure 6). At 3000 and 3600 m, we found positive relationships (p < 0.05) between A. fabri H trees’ growth indices and previous December precipitation (Figure 6a). Conversely, negative correlations were recorded between the growth of A. fabri H trees and previous July precipitation at 3000 m, as well as between A. fabri H growth and current August precipitation at 3600 m. In contrast, no significant positive correlations were observed for A. fabri D trees with precipitation across all elevation levels (Figure 6b). However, a significant negative correlation was found between the growth of A. fabri D trees and previous March and July precipitation at 3600 m, along with current August precipitation at the same elevation.

Analyses of correlations with temperature indicated that A. fabri H trees displayed significant positive relationships with mean temperature in the previous January, April, and December, as well as with current October temperature at 3600 m (Figure 6a). Conversely, significant negative correlations were noted between A. fabri H trees’ growth indices and the previous February, August, and October at 3000 m. For A. fabri D trees, significant positive correlations were found with the previous April and November temperatures at 3600 m, along with the current July temperature at both 3000 and 3600 m, and also with August and October temperatures at 3600 m (Figure 6b).

Regarding SPEI, only A. fabri H trees exhibited a significant positive correlation with SPEI in the current June at 3000 m (Figure 6a). Furthermore, analyses of the relationships with VPD revealed that A. fabri H trees showed significant negative correlations with the mean VPD in the preceding June at both 2800 m and 3000 m (Figure 6a). In contrast, for A. fabri D trees, significant negative correlations were identified with the preceding June and July VPD at 2800 m, as well as with the previous June VPD at 3000 m (Figure 6b).

3.5. Soil and Leaf Nutrient Concentrations

Soil-available N concentrations varied across elevations (p < 0.05), with the highest values found at 3600 m (39.2 mg g⁻¹) (Figure S7). Furthermore, a significant difference (p < 0.001) in soil AP concentrations between elevations was noted, with values peaking (59.26 mg g⁻¹) at 3600 m (Figure S7). In contrast, soil DOC concentrations did not show significant mean differences across elevations (Figure S7).

Leaf N and C concentrations were significantly lower (p < 0.001) in A. fabri D than in H trees, regardless of elevation (Figure S7). Specifically, the leaf N concentration for A. fabri H trees ranged from 9.27 to 9.32 mg g⁻¹, whereas those of D trees ranged from 6.09 to 6.19 mg g⁻¹ (Figure S7). In the case of leaf P concentrations, differences were found among elevations decreasing upwards, but not between vigor classes (Figure S7).

4. Discussion

The relationships between the health of A. fabri trees and their environmental conditions, including climate variables, biotic stressors, and soil conditions, provide crucial insights into why some trees are healthy while others are in decline. Over the period from 1950 to 2019, precipitation decreased at 2800 m, whereas temperature increased across all elevations. Notably, the relationship between temperature and precipitation was negative at 2800 m, indicating a higher drought stress as the climate continues to warm. Tree growth patterns highlighted a marked disparity, with H trees outperforming D trees at 2800 m. Furthermore, bark beetle infestations were common in D trees, peaking at the highest elevation.

4.1. Elevation Changes in Climate Trends and Their Role in Tree Health

Variations in climatic parameters along the elevational gradient significantly influence tree growth and post-drought recovery [68]. Over the period from 1950 to 2019, a notable decline in precipitation at 2800 m suggests intensified drought stress affecting A. fabri growth and health. However, despite this reduction in precipitation, growth trends for both the H and D trees have generally increased since the 1980s, indicating a more nuanced response to climate change. Notably, the reduction in precipitation during the early 2000s, combined with marked seasonal temperature increases (Figure S1), may have exacerbated drought stress for the declining trees at 2800 m. Yet, the warming trends at mid and higher elevations likely enhanced physiological performance and extended growing seasons during critical periods, supporting sustained growth in healthy trees and partially offsetting the negative effects of drought. This complexity aligns with established knowledge that water availability strongly influences tree health, where drought stress can reduce resilience and lead to mortality in sensitive individuals [68,69,70]. Simultaneously, warming-induced climatic stability at higher elevations can promote growth in healthier trees. Thus, the observed increases in growth alongside drought signals underscore species-specific and elevation-dependent interactions between precipitation and temperature. Such dual effects are consistent with global observations in wet montane ecosystems, where drought and warming interact to influence tree health outcomes [71]. Overall, the interaction of elevation, precipitation decline, and temperature rise shapes the health trajectories of A. fabri, highlighting the importance of considering multiple climatic drivers when assessing species’ responses to changing environments.

4.2. Tree Growth Trends and Climate–Growth Variability

The growth trends underscore the resilience of healthy trees across all elevations. Notably, at 2800 m, declining trees experienced substantial suppressions from the 1950s to 1970s (Figure 5a), consistent with prolonged VPD periods and associated stress (Figure 2h). This suppression was followed by partial recovery, likely attributed to climatic amelioration during the 1990s. At 3000 m and 3600 m, most trees were healthy and showed little suppression during the late 20th century (Figure 5b,c), suggesting a better capacity to withstand harsher climatic conditions. The recent increases in BAI of healthy trees at all elevations indicate sustained resilience mechanisms that merit further investigation. These findings resonate with studies on alpine conifers, where tree vigor correlates with the ability to endure climatic extremes such as droughts [2,72]. Various studies have also shown the decline in alpine forests [73,74].

Despite the notable increase in the number of disturbances after 2000 (Figure 5b,c), the BAI of Abies fabri trees also shows a marked rise during the same period (Figure 4). This apparent discrepancy could be attributed to several non-mutually exclusive factors. First, increased BAI may reflect a recovery phase following earlier growth suppression events, where surviving trees experience release due to reduced competition after disturbance-induced mortality. Second, early stages of biotic disturbances, such as bark beetle infestations, may selectively affect weaker individuals, inadvertently favoring residual tree growth through resource reallocation [47]. Lastly, the lagged physiological responses to earlier climatic or biotic stressors may contribute to complex growth dynamics, making the BAI response not linearly correlated with disturbance frequency [75].

The climate–growth relationship highlights significant differences between healthy and declining trees. Healthy trees exhibit positive correlations with prior-year precipitation, particularly in December, which is not observed in declining trees. This suggests that healthy trees are better able to utilize available soil moisture before cambial onset in spring. Conversely, the negative correlations with precipitation during summer months suggest that excessive water during this period may impede growth, potentially due to hypoxic soil conditions, low radiation levels, or fungal infections during the monsoon season. In such anaerobic environments, root respiration is hindered, leading to stress responses in trees, such as reduced nutrient uptake, slowed growth rates, or even tree decline [76,77]. Declining trees, however, showed no significant positive correlations with precipitation, underlining their diminished responsiveness to climate variability.

The role of VPD adds another layer to understanding these dynamics. The significant negative correlations observed between VPD and tree growth in healthy trees, particularly during the preceding June at 2800 m and 3000 m (Figure 6), suggest that high VPD exacerbates water stress. Elevated VPD levels lead to increased transpiration rates, intensifying water loss and limiting the availability of soil moisture, thereby affecting growth [24,25,78]. In our study, declining trees also exhibited significant negative correlations with VPD during the preceding summer months at 2800 m and 3000 m. However, their overall diminished growth responsiveness highlights an already compromised physiological state, making them less capable of adapting to high evaporative demand. These findings align with previous studies that emphasize the detrimental effects of high VPD on forest productivity, particularly under water-limiting conditions, where tree mortality and growth suppression are linked to increased VPD and reduced water use [79,80]. Furthermore, positive correlations between growth and winter/spring temperatures in healthy trees at 3600 m suggest that warmer temperatures facilitate growth at high elevations and can promote upward shifts in the study species [79].

4.3. Disturbance Dynamics and Their Links with Tree Health

Declining trees experienced a period of growth suppression from the 1950s to the late 1970s, particularly at low elevation (Figure 5). These findings align with existing literature suggesting that trees experiencing adverse climatic conditions are more susceptible to disturbances from biotic factors such as beetle infestations [29,70]. Drought has been strongly associated with historic pest outbreaks in dry regions [81]. However, its effect varies with the feeding guilds of insects, the substrate they feed on, and the intensity of drought [82]. In our study, bark beetle infestations were significantly higher in declining than in healthy trees (Figure 3c) at 2800 and 3000 m, where drought stress was higher. This aligns with the understanding that trees weakened by drought and other abiotic stresses are less capable of defending themselves against insect attacks, leading to higher mortality rates in already compromised trees [29,70]. While precipitation decreased at 2800 m, drought stress at 3000 m was primarily driven by the significant increase in vapor pressure deficit (VPD) as shown in Figure 2h. This distinction provides a more accurate and nuanced explanation of the differing environmental factors contributing to drought stress across the elevations. Therefore, our data suggest positive feedback where the underperformance of drought-stressed trees may increase their vulnerability to pests, thereby exacerbating the overall decline and contributing to tree death. We therefore suggest that more research focusing on reconstructing bark beetle attacks should be performed in the study region using satellite imagery information combined with tree-ring data. These methodologies can provide high spatial and temporal resolutions, enabling researchers to monitor forest disturbances more effectively and detect early signs of decline and infestation.

It is also crucial to note the distinct patterns observed at 3000 m elevation, where trees exhibited the highest mean DBH and height values, suggesting optimal growing conditions for A. fabri at this mid-elevation site. Despite this, A. fabri D trees at 3000 m still showed periods of significant growth suppression in the 1960s and 2000s, indicating that even in generally favorable conditions, these trees experienced vulnerability to environmental stressors. This contrasts with the patterns at 2800 m, where suppressions were recorded by both healthy and declining trees in the 1970s, and at 3600 m, where suppressions broadly affected trees from 1964 to 1978 and again from 1999 to 2004. These elevation-specific differences highlight a nuanced response to disturbance dynamics, where the health status and climatic conditions at 3000 m present a unique context for understanding tree resilience and susceptibility to bark beetle infestations.

4.4. Climate Change and Its Impacts on a. Fabri Health

The ongoing effects of climate change present significant challenges to tree species adapted to specific climatic conditions [83,84,85]. Although the growth of D A. fabri trees has increased since the 1980s, particularly at 2800 m, as shown in Figure 2, this apparent growth enhancement may mask underlying physiological stresses caused by elevated temperatures and reduced soil moisture availability. Such stressors can exacerbate drought vulnerability and compromise long-term tree health, even when short-term growth responds positively, possibly due to extended growing seasons or acclimation processes.

In addition to temperature increases, acid rain may play a compounding role in the decline of trees’ health [86,87]. The potential deposition of acid rain in the GM region, driven by industrial emissions and climate-driven changes in precipitation patterns, could lead to soil acidification, nutrient leaching, and disruption of nutrient uptake processes. Such impacts are likely to affect tree growth and resilience, particularly for declining populations already stressed by other climatic and environmental factors. Acid rain could further contribute to the reduced N and P concentrations observed in declining trees, potentially weakening their defense mechanisms against pests and environmental stressors [88]. Research on the presence and impacts of acid rain in the GM region remains absent, although such studies have been conducted on nearby mountainous systems. For instance, [89] developed a Lagrangian backward trajectory model to assess acid rain effects in the Emei Mountainous, southwest China. His findings revealed significant contributions from distant pollutant sources in the Sichuan Basin, especially along the northeast direction of Emei. These results underscore the potential for similar atmospheric processes to affect GM, given its proximity and shared climatic influences with Emei Mountain. Conducting similar studies in GM is essential to determine the presence, sources, and impacts of acid rain on A. fabri and other forest ecosystems.

In addition to abiotic stressors, biotic factors, such as root pathogens, may further exacerbate the decline of A. fabri. Root pathogens like Heterobasidion species are known to infect coniferous trees, causing root and butt rot that compromise water transport, nutrient uptake, and structural stability [90,91]. These pathogens often exploit trees already weakened by environmental stressors such as drought or nutrient deficiencies. Declining populations of A. fabri, particularly at lower elevations, may be more susceptible to such infections due to their reduced nutrient concentrations and physiological stress. While no direct evidence of Heterobasidion infections has been documented for A. fabri in GM, the potential role of root pathogens in accelerating tree decline should not be overlooked, especially in the context of multi-stressor environments.

The decline in A. fabri abundance with elevation reflects the interplay of various stressors associated with changing climatic conditions. Lower elevations experience more severe and prolonged droughts due to rising temperatures, exacerbating soil nutrient depletion and reducing tree resilience. In these areas, declining trees may also face greater competition from other species thriving under relatively favorable conditions, further limiting resource availability. Conversely, higher elevations witness reduced competition, allowing healthy trees to capitalize on available resources and maintain stronger defenses against pests and pathogens. However, intermediate elevations showed the highest mean DBH and height values, and BAI peaked at low elevations, suggesting that tree-to-tree competition may play a less significant role in this case. The combined impacts of climate warming, acid rain, and potential root pathogens underscore the complexity of stressors affecting A. fabri. Comprehensive research into the presence and role of Heterobasidion and similar pathogens in GM is critical to elucidate their contributions to tree health dynamics. Developing integrated management strategies to mitigate these biotic and abiotic stresses is vital for conserving A. fabri populations and ensuring the resilience of these forest ecosystems under future climate scenarios. The combined impacts of the study stressors may be enhancing an upward shift in A. fabri along the GM elevational gradient. Establishing a plot network in combination with forest inventory data and tree-ring reconstructions could help to elucidate this compositional shift. This is corroborated by Hernández et al. [92], who found evidence of an upward expansion of Abies alba in the subalpine forests of the Spanish Pyrenees, characterized by increased recruitment, higher growth rates, and lower mortality, in contrast to declining regeneration, vitality, and rising mortality in montane populations. Their study highlights how marginal rear-edge populations are experiencing both contractions and expansions in response to climate aridification and biotic interactions, emphasizing the importance of elevation-specific monitoring to inform conservation strategies.

5. Conclusions

The current study underscores the significance of understanding the multifaceted influences on tree health and growth within the context of climate change. While reduced precipitation was observed only at lower elevations (2800 m), the increasing proportion of declining trees at higher elevations suggests that factors beyond water stress, particularly bark beetle infestations, physiological limitations, and possibly acid rain or pathogen pressures, play a more dominant role in driving decline at upper sites. This highlights that elevation-dependent tree health dynamics in the GM are governed by a complex interplay of biotic and abiotic stressors, rather than by precipitation trends alone. The relatively mild and wet conditions at mid and high elevations create a more favorable environment for growth in combination with higher soil N concentrations. In contrast, declining trees showed lower N and C leaf concentrations. Future conservation efforts of A. fabri low-elevation forests must integrate a comprehensive understanding of the interactions of multiple stress factors to mitigate their adverse effects. The final aim is to promote healthier tree populations capable of withstanding future climate-related challenges, including more frequent and severe droughts. By enhancing forest resilience and embracing adaptive management strategies, we can try to bolster the recovery of declining A. fabri stands.