Altitudinal Difference of Growth–Climate Response Models in the Coniferous Forests of Southeastern Tibetan Plateau, China

: Characterized as a climatologically sensitive region, the southeastern Tibetan Plateau (STP) is an ideal location for dendrochronological research. Here, five tree-ring width (TRW) chronologies were developed: three for Picea likiangensis along altitudinal gradients from 3600 to 4400 m a.s.l. and two for Sabina saltuaria and Abies squamata from 4200 m a.s.l. Significant differences in the growth rates and age composition of Picea likiangensis were observed at various elevation gradients. The chronology statistics (mean sensitivity, etc.) fluctuated with the elevation gradient. Picea likiangensis showed distinct growth patterns in response to climatic variability along the altitude gradient: the minimum temperature influenced tree growth at lower and middle altitudes, while higher altitudes were affected by precipitation. The radial growth of different tree species growing in the same region is controlled by the same climatic factors. Sabina saltuaria and Abies squamata exhibited similar growth responses to Picea likiangensis . Stand conditions (wind speeds, slope, and elevation) and biotic factors (the depth of root, forest type, tree age, and sensitivity) can partially explain why the ring width–climate relationships change with altitude.


Introduction
Human influence has warmed the climate at an unprecedented rate in at least the last 2000 years, particularly affecting high-mountain regions where the climate changes more rapidly than in surrounding lowlands [1,2].Climate change will affect the population structure, distribution, and annual growth rates of tree species [3][4][5][6][7].In alpine areas, the temperature-precipitation (hydrothermal) coupling pattern alters with increasing altitude, as altitude corresponds to a natural cooling and humidification gradient [7][8][9].Temperature and precipitation are the primary drivers of tree growth [10,11].Consequently, plants exhibit varied adaptations in growth, survival, and metabolism in response to altitude changes, and the relationship between tree growth and climate often varies with altitude [12,13].Therefore, a comprehensive understanding of the growth-climate responses of high-elevation tree species across their distribution range is essential for devising appropriate forest management and conservation strategies to mitigate the adverse impacts of climate change [14].
Predicting changes in growth dynamics and evaluating the impact of climate change requires a deeper understanding of tree growth responses along elevation gradients.Throughout the altitudinal range of tree species, the lower and upper distribution limits are identified as critical zones that are particularly sensitive to local climate changes [15,16].It is

Study Area
The study was conducted in the high-elevation forest ecosystems of the STP, and the sampling sites were located in the Haizi mountain, China (99 • 56 ′ -99 • 59 ′ E, 29 • 26 ′ -29 • 29 ′ N) (Figure 1).This area exhibits distinct altitudinal belts and vegetation types with a high level of plant endemism due to its topographical complexity [39].Various climate systems, including the eastern Asian monsoon, the Indian monsoon, and the continental westerlies, play an active role in driving the regional climate patterns of this region [40].Based on the data acquired from a local meteorological station (Daocheng, 29.03 • N, 100.18 • E, 3728 m a.s.l.), the annual precipitation is 660 mm, most of which falls between June and September.The annual mean temperature is 4.58 • C, with a mean temperature of 12.2 • C in July and −4.9 • C in January (Figure 2).According to the climate data over the past six decades at Daocheng station, significant warming trends have been observed in annual mean temperatures, but total annual precipitation did not show a clear changing trend (Figure 3).

Study Area
The study was conducted in the high-elevation forest ecosystems of the STP, and the sampling sites were located in the Haizi mountain, China (99°56′-99°59′ E, 29°26′-29°29′ N) (Figure 1).This area exhibits distinct altitudinal belts and vegetation types with a high level of plant endemism due to its topographical complexity [39].Various climate systems, including the eastern Asian monsoon, the Indian monsoon, and the continental westerlies, play an active role in driving the regional climate patterns of this region [40].Based on the data acquired from a local meteorological station (Daocheng, 29.03° N, 100.18°E, 3728 m a.s.l.), the annual precipitation is 660 mm, most of which falls between June and September.The annual mean temperature is 4.58 °C, with a mean temperature of 12.2 °C in July and −4.9 °C in January (Figure 2).According to the climate data over the past six decades at Daocheng station, significant warming trends have been observed in annual mean temperatures, but total annual precipitation did not show a clear changing trend (Figure 3).

Study Area
The study was conducted in the high-elevation forest ecosystems of the STP, and the sampling sites were located in the Haizi mountain, China (99°56′-99°59′ E, 29°26′-29°29′ N) (Figure 1).This area exhibits distinct altitudinal belts and vegetation types with a high level of plant endemism due to its topographical complexity [39].Various climate systems, including the eastern Asian monsoon, the Indian monsoon, and the continental westerlies, play an active role in driving the regional climate patterns of this region [40].Based on the data acquired from a local meteorological station (Daocheng, 29.03° N, 100.18°E, 3728 m a.s.l.), the annual precipitation is 660 mm, most of which falls between June and September.The annual mean temperature is 4.58 °C, with a mean temperature of 12.2 °C in July and −4.9 °C in January (Figure 2).According to the climate data over the past six decades at Daocheng station, significant warming trends have been observed in annual mean temperatures, but total annual precipitation did not show a clear changing trend (Figure 3).

Sample Collection and Dendrochronological Analyses
The hillside near the Dengpo Township, Daocheng County, China was chosen as the sampling location; the elevation of the sampled P. likiangensis forest ranges from 3550 to 4450 m a.s.l.P. likiangensis is a shallow-rooted species that can tolerate cold weather [37].Consequently, it is widely distributed on the north-facing slopes of high mountains where the climate is cold and humid.Though P. likiangensis is the dominant tree species on this hillside, the community structure at each sampling site varies significantly with elevation.In the high-elevation zone, P. likiangensis and S. saltuaria form an open mixed forest, with Rhododendron decorum Franch and Rhododendron phaeochrysum present in the understory.The forest in the middle elevation range is an open stand primarily dominated by P. likiangensis, with a few R. phaeochrysum and Salix sclerophylla in the understory.In the lower elevation zone, the forest remains characterized by an open, pure stand, predominantly composed of P. likiangensis, and accompanied by some Salix sclerophylla.
According to the standard dendrochronological techniques for sample preparation and chronology development [41][42][43], P. likiangensis trees were sampled from three elevational belts at the species' upper distribution limit (4250-4350 m a.s.l., DPH), at the middle range (3850-3950 m a.s.l., DPM), and at the lower distribution limit (3650-3750 m a.s.l., DPL).S. saltuaria trees were sampled at the site corresponding to DPH (4170-4283 m a.s.l., DPF).In addition, we sampled some Abies squamata Mast.from another hillside near the DPH (4355-4398 m a.s.l., ZWH) (Figure 1).At each sampling site, two to three increment cores per tree were collected from 20 to 26 mature and heathy trees using the increment borer of a 5.15 mm diameter (Supplemental Figure S1).Specifically, to develop chronologies, 47, 56, 50, 44, and 42 increment cores were collected at the breast height from the DPH, DPM, DPL, DPF, and ZWH sites, respectively (Table 1).The increment cores were stored in paper tubes to prevent damage.
The preparation and processing of samples were carried out following international general specifications and standards [41,42,44].First, the samples were dried and fixed in wooden slots; then, we used consecutively finer grades of sandpapers to smooth the core surfaces until the ring boundaries and cells were clearly visible under a microscope.We used skeleton plots for cross-dating and the TSAP-win TM (v4.81c) standard annual ring analysis software for auxiliary cross-dating.Tree-ring widths were measured with a LinTAB 6 system (Rinntech, Heidelberg, Germany) at a resolution of 0.001 mm.Each core was measured at least twice.Then, COFECHA software (2012) [44,45] was utilized to check the results of cross-dating and corrected errors following the microscopic

Sample Collection and Dendrochronological Analyses
The hillside near the Dengpo Township, Daocheng County, China was chosen as the sampling location; the elevation of the sampled P. likiangensis forest ranges from 3550 to 4450 m a.s.l.P. likiangensis is a shallow-rooted species that can tolerate cold weather [37].Consequently, it is widely distributed on the north-facing slopes of high mountains where the climate is cold and humid.Though P. likiangensis is the dominant tree species on this hillside, the community structure at each sampling site varies significantly with elevation.In the high-elevation zone, P. likiangensis and S. saltuaria form an open mixed forest, with Rhododendron decorum Franch and Rhododendron phaeochrysum present in the understory.The forest in the middle elevation range is an open stand primarily dominated by P. likiangensis, with a few R. phaeochrysum and Salix sclerophylla in the understory.In the lower elevation zone, the forest remains characterized by an open, pure stand, predominantly composed of P. likiangensis, and accompanied by some Salix sclerophylla.
According to the standard dendrochronological techniques for sample preparation and chronology development [41][42][43], P. likiangensis trees were sampled from three elevational belts at the species' upper distribution limit (4250-4350 m a.s.l., DPH), at the middle range (3850-3950 m a.s.l., DPM), and at the lower distribution limit (3650-3750 m a.s.l., DPL).S. saltuaria trees were sampled at the site corresponding to DPH (4170-4283 m a.s.l., DPF).In addition, we sampled some Abies squamata Mast.from another hillside near the DPH (4355-4398 m a.s.l., ZWH) (Figure 1).At each sampling site, two to three increment cores per tree were collected from 20 to 26 mature and heathy trees using the increment borer of a 5.15 mm diameter (Supplemental Figure S1).Specifically, to develop chronologies, 47, 56, 50, 44, and 42 increment cores were collected at the breast height from the DPH, DPM, DPL, DPF, and ZWH sites, respectively (Table 1).The increment cores were stored in paper tubes to prevent damage.The preparation and processing of samples were carried out following international general specifications and standards [41,42,44].First, the samples were dried and fixed in wooden slots; then, we used consecutively finer grades of sandpapers to smooth the core surfaces until the ring boundaries and cells were clearly visible under a microscope.We used skeleton plots for cross-dating and the TSAP-win™ (v4.81c) standard annual ring analysis software for auxiliary cross-dating.Tree-ring widths were measured with a LinTAB 6 system (Rinntech, Heidelberg, Germany) at a resolution of 0.001 mm.Each core was measured at least twice.Then, COFECHA software (2012) [44,45] was utilized to check the results of cross-dating and corrected errors following the microscopic examination of tree-ring characteristics [46].We applied a negative exponential regression function and a 67% of the series length cubic spline function to remove the age-related trend for each series using the ARSTAN program (2008) [46].The analysis results revealed that the detrending methods had a negligible impact on the quality of the chronologies (Supplemental Tables S1 and S2).The bi-weight robust estimation of the mean was employed for all the series to produce the standard chronology (STD), residual chronology (RES), and ARTSAN chronology (ARS).The mean sensitivity (M.S.), standard deviation (S.D.), and first-order autocorrelation coefficient (AC1) of the chronology were calculated to evaluate the quality of the chronology.In addition, the signal-to-noise ratio (SNR) and first principal component variance interpretation (PC1) were also calculated.The reliability and signal strength of each standard chronology was assessed by a 50-year moving expressed population signal (EPS) and the mean series inter-correlations (Rbar) [10].

Climate Data
The instrumental data, including monthly total precipitation (P), mean temperature (T mean ), mean minimum temperature (T min ), and mean maximum temperature (T max ) from the previous May to the current October, were collected from the local meteorological station (Daocheng, 29.03 • N, 100.18 • E, 3728 m a.s.l.), which is located 52 km from our sampling sites (Figure 1), with coverage from A.D. 1959 to 2021 (Data from https://data.cma.cn/, accessed on 24 April 2023).This study employed the moving t-test to detect mean shifts in the temperature and precipitation time series [10].The results indicate that there are no abrupt changes in temperature and precipitation series at the Daocheng station.Consequently, the meteorological data demonstrate good homogeneity and can be reliably used for subsequent analysis and research.The Standardized Precipitation-Evapotranspiration Index (SPEI) [47] was used to derive regional moisture conditions.The SPEI dataset was obtained from CRUTS3.23 (http://sac.csic.es/spei,accessed on 24 April 2023).Climate-growth relationships were determined by using Pearson's correlation analyses.Correlation analyses between the tree-ring width index (TRI) of each residual chronology and climatic variables were performed for 18 months in total, starting in May of the previous growing season, and ending in October of the current growing season.

Tree-Ring Width Chronologies and Statistics
There were significant differences in the growth rates and age composition among chronologies of various elevation gradients.The age of P. likiangensis trees increased with elevation.The average age of trees across all P. likiangensis stands ranged from 108 to 145 years; the youngest tree (22 years old) was found at the lower elevation stand (DPL) while the oldest tree (389 years old) was found at the high-elevation site (DPH).Because many of the standing trees with large diameters at the DPH are rotten inside, the length of the chronologies does not accurately reflect the age of the old-growth fir forests at the DPH.The average annual radial growth rate was found to decrease with increasing age at the DPL; however, it did not change significantly at the DPM and DPH (Figure 4).
To retain more signals [48], the STD chronologies, established by the negative exponential regression function detrending method, were selected for subsequent analysis (Figure 5).The correlation analysis of the STD chronologies of P. likiangensis growing at three different elevations showed that the TRI variation at the DPL and DPM had a high degree of agreement, with correlations exceeding the 99% significance level (r = 0.673).In contrast, the TRI variation at the DPH differed significantly from those at both the DPM and DPL, with lower correlations (r = 0.171 and r = 0.149, respectively, p > 0.01) (Table S1).This phenomenon indicated that the growth of P. likiangensis at the DPL and DPM is likely governed by similar climatic factors due to the consistent TRI observed in trees at the DPL and DPM.In contrast, radial growth at the DPH appears to be influenced by climatic factors distinct from those affecting the DPL and DPM.
than that of the high altitudes and the low altitudes (Table 2).The AC1 values for all the sampling sites are high, indicating that there are strong climatic lag-effects [10].Despite the differences in the chronological characteristics of the data at different elevations, all the above data indicated that the P. likiangensis growing at different elevations in the study area contain more environmental information and can be used for dendroclimatological research.As to the other species, the EPS of them was below the threshold of 0.85 (Table 2), which showing that there was no reliable signal contained in S. saltuaria and A. squamata chronologies.The M.S., SNR, and EPS of the P. likiangensis were higher than those of S. saltuaria and A. squamata, indicating that the annual ring index of P. likiangensis contains more environmental information and is more suitable for the study of dendrochronology.T/C: core/tree in the common interval, M.S.: mean sensitivity, SNR: signal-to-noise ratio, EPS: expressed population signal, WTR: mean correlations within-trees, BTR: mean correlations between trees, and Rbar: mean series inter-correlations.

Response to Climatic Condition
Given the lag-effect of the climate on radial tree growth, we selected meteorological data from the beginning of the previous year's growing season to the end of the current year's growing season (previous May to current October) for dendroclimatic analysis.The tree radial growth of P. likiangensis showed positive correlations with mean temperature and mean minimum temperature at the DPL and DPM during the winter and spring It can be observed that most statistical parameters (including M.S., EPS, SNR, WTR, and BTR) at the DPM exhibited a higher value than the other two sites (DPH, DPL), indicating that coherence among the tree-ring width series at the middle altitudes are better than that of the high altitudes and the low altitudes (Table 2).The AC1 values for all the sampling sites are high, indicating that there are strong climatic lag-effects [10].Despite the differences in the chronological characteristics of the data at different elevations, all the above data indicated that the P. likiangensis growing at different elevations in the study area contain more environmental information and can be used for dendroclimatological research.As to the other species, the EPS of them was below the threshold of 0.85 (Table 2), which showing that there was no reliable signal contained in S. saltuaria and A. squamata chronologies.The M.S., SNR, and EPS of the P. likiangensis were higher than those of S. saltuaria and A. squamata, indicating that the annual ring index of P. likiangensis contains more environmental information and is more suitable for the study of dendrochronology.

Response to Climatic Condition
Given the lag-effect of the climate on radial tree growth, we selected meteorological data from the beginning of the previous year's growing season to the end of the current year's growing season (previous May to current October) for dendroclimatic analysis.The tree radial growth of P. likiangensis showed positive correlations with mean temperature and mean minimum temperature at the DPL and DPM during the winter and spring months, and it showed positive correlations with the SPEI and negative correlations with the mean maximum temperature at the DPH during the summer months.The results regarding the TRI's response to climatic factors is shown in Figure 6.
On the monthly scale, the DPL and DPM indicated that the growth of trees is significantly positively responsive to most of the temperature parameters (T mean , T min ) during the previous and last winter (Dec., Jan., Feb.) and spring (Mar., Apr., May) months, although the magnitude and intensity of the correlation coefficients varied; they were all significant at the 99% confidence level.In contrast, the influence of precipitation on the growth of P. likiangensis at the DPL and DPM was minimal.The DPL showed a positive correlation with precipitation in the current June and October, but the correlation coefficients were not significant (Figure 6a), and the DPM did not show any relationships with precipitation from the previous and current year.In addition, the DPL and DPM relationship with the SPEI only showed a positive correlation in August, but the significance did not reach the 99% confidence level (Figure 6b).For the high-altitude sites, the DPH showed significantly positive responses to precipitation in the early growing season (from May to July) at the 99% confidence level (Figure 6c).
The results of the Pearson correlation analyses conducted on seasonal and annual scales indicate a significant positive correlation between radial tree growth and T min across all time periods, both during the growing season and throughout the year.The highest correlation between the DPL and temperature parameters occurred in the previous November to the current August (P Nov -C Aug ), especially for T min (r = 0.560, p < 0.01) (Figure 6a, Supplemental Figure S2), and the same results were also found for the DPM (r = 0.600, p < 0.01) (Figure 6b, Supplemental Figure S3).The correlation coefficient values for precipitation were significantly lower than those for temperature, although the DPL showed positive correlations with precipitation from the previous October to the current July (r = 0.356, p < 0.01) (Supplemental Figure S2), and the DPM showed strong positive correlations with precipitation from the previous May to the current April (r = 0.321, p < 0.01) (Supplemental Figure S3).When it comes to the high-altitude site, the DPH showed a significantly positive correlation with the SPEI (P May -P Dec , r = 0.341, p < 0.01) and precipitation (P Oct -C Jun , r = 0.394, p < 0.01) (Figure 6c, Supplemental Figure S4).

Climate-Growth Relationships along the Altitudinal Gradient
All detected correlations between the radial growth of P. likiangensis and climate data were not significantly associated with the elevation gradients; thus, the influence of altitude on tree growth was not uniform in the STP.In our research, both temperature and precipitation were identified as main factors affecting tree growth, with effects varying Though the EPS of S. saltuaria and A. squamata chronology did not reach the level of 0.85, we still analyzed the relationship between the TRI and climate factors to verify whether the radial growth of different tree species growing in the same region is controlled by the same climatic factors.At the DPF, tree growth showed a significant positive correlation with precipitation in the previous October, whereas a significant negative correlation with T max was observed in the current May (r = 0.355, p < 0.01) (Figure 6e).At the seasonal and annual scale, the DPF has a positive correlation with the precipitation in the previous October to the current February (P Oct -C Feb , r = 0.308, p < 0.05) (Figure 6e, Supplemental Figure S5).As for ZWH, the result of the correlation analysis showed that the radial growth of A. squamata is limited by the precipitation during the current September (r = 0.393, p < 0.01).At the seasonal and annual scale, ZWH has a positive correlation with the precipitation during the previous October to the current February (P Oct -C Feb , r = 0.308, p < 0.05) (Figure 6d, Supplemental Figure S6).

Climate-Growth Relationships along the Altitudinal Gradient
All detected correlations between the radial growth of P. likiangensis and climate data were not significantly associated with the elevation gradients; thus, the influence of altitude on tree growth was not uniform in the STP.In our research, both temperature and precipitation were identified as main factors affecting tree growth, with effects varying according to altitude and tree species.Similar results were also found in previous studies [9,19,[49][50][51].At higher elevations, the radial growth of P. likiangensis was mainly limited by precipitation, whereas at middle and lower elevations, temperature played a dominant role in ring width formation.This finding contradicts the general principle of limiting factors, according to which it is generally believed that a tree species is mainly affected by temperature at their upper altitudinal limits and by precipitation at their lower limits [17,18,51,52].However, similar phenomena have been observed in some other areas of the STP, and these phenomena can be explained by the knowledge of tree physiology.
The T min from the previous November to the current August (P Nov -C Aug ) was a dominant factor influencing the radial growth of P. likiangensis at lower and middle altitudes.This climatic response model was typical in many areas of the central Hengduan Mountains (HM) [53], the STP [54][55][56], and the central Himalayas [14,22].The same climatic response model was also found in other regions [57].For example, at Balang mountain in the Wolong Natural Reserve (Western Sichuan Province, China), the radial growth of Picea brachytyla var.complanate showed a general positive correlation with the temperature of all months from November of the prior growth year to October of the current growth year [53].The study conducted by Keyimu et al. [56] also showed that the annual T min was the key climatic factor influencing radial tree growth in conifer species at 98 sites in the STP.So, the significant positive correlation between P Nov -C Aug T min and local tree growth is reasonable.During the summer, T min influences the division and enlargement of cells and influences xylem lignification [58,59].Thus, T min , which is frequently measured at night, is significant for radial tree growth because xylem lignification occurs primarily at night [56].Low summer soil temperature can limit the growth of roots and their function in water uptake, and hence limit the tree growth in our study area.A cold August may lead to an early cessation of tree growth and a reduction in the rate of cell division, consequently reducing tree-ring widths.In contrast, a warm August may lead to a longer growing season and may still allow the formation of latewood cells [10].For instance, Zhu et al. [39] found that low temperatures in August were the primary factor limiting the growth of P. likiangensis, which grew in the eastern part of the Nyainqentanglha Mountains.A positive effect of the summer minimum temperature on tree growth was also reported for the entire TP [60][61][62][63].A warm climate in the spring has a significantly positive effect on tree growth; higher spring temperatures can lengthen the growing season by breaking dormancy in advance, boosting the advanced division of tree cambium cells [24,64], and increasing leaf area to increase the photosynthetic rate [65].It is found that the tree growth in the surrounding areas of our study site were also positively limited by spring temperature [33,51].A warm winter may protect the needles from frost damage, prevent root damage caused by freezing embolism, and maintain the roots in a photosynthetically active state for the next year [66].However, in a cold winter, the soil from a deep layer of frost can delay the time of thawing and shorten the growing season, resulting in the formation of narrow rings [14,67].This can explain why the TRI in the DPL and DPM show a notable positive correlation with the last winter.The positive correlation between radial growth and winter temperatures in the central HM was also supported by previous studies [23,34,68].
In contrast, at high elevations, the radial growth of P. likiangensis was mainly limited by precipitation during the period between the end of the last growing season and the beginning of this growing season.Generally, a tree species is mainly affected by temperature at its upper altitudinal limit since the temperature drops at a rate of 0.6 • C per 100 m of elevation gain as the altitude rises.However, in our study region, the radial growth of P. likiangensis, S. saltuaria, and A. squamata were all positively associated with precipitation and the SPEI, and negatively associated with the maximum temperature.Gou et al. [69] found that the tree growth of Abies forrestii Coltm., which grew in Xiangcheng, central HM, was also significantly positively associated with precipitation during the previous September to the current June.This phenomenon was also found in the central Himalayas.Liang et al. [70] found that in the central Himalayas, precipitation decreases with increasing elevation.Thus, trees growth at high altitudes in this area was mainly limited by moisture stress rather than temperatures [22].But in our research, according to the data acquired from automatic weather stations (HOBO U30, USA), on some days the relative humidity (RH) in the DPL was higher than that in the DPH (Figure 7).So, the question of why the growth of P. likiangensis, S. saltuaria, and A. squamata at high altitudes was mainly constrained by precipitation and the SPEI needs further study.Therefore, we discuss some possible reasons in the next section.In the future, further monitoring is needed to test the reliability of this conclusion.increasing elevation.Thus, trees growth at high altitudes in this area was mainly limited by moisture stress rather than temperatures [22].But in our research, according to the data acquired from automatic weather stations (HOBO U30, USA), on some days the relative humidity (RH) in the DPL was higher than that in the DPH (Figure 7).So, the question of why the growth of P. likiangensis, S. saltuaria, and A. squamata at high altitudes was mainly constrained by precipitation and the SPEI needs further study.Therefore, we discuss some possible reasons in the next section.In the future, further monitoring is needed to test the reliability of this conclusion.

Factors May Influence Growth Response to Climate Change
Although there are now many studies that have reached conclusions which do not agree well with the general principle of limiting factors, none of them have explained the reasons for this difference.Therefore, in this paper, we grouped possible reasons into two categories.The first possible reason concerns the stand conditions (wind speeds, slope, elevation, etc.).The altitude of our research area ranged from 3600 to 4400 m, which is much higher than other study sites; thus, the temperature is lower than that in other study areas, which caused the tree growth in a relative lower elevation area to show a significant positive relationship with the temperature (DPL, DPH).The increase in the elevation and

Factors May Influence Growth Response to Climate Change
Although there are now many studies that have reached conclusions which do not agree well with the general principle of limiting factors, none of them have explained the reasons for this difference.Therefore, in this paper, we grouped possible reasons into two categories.The first possible reason concerns the stand conditions (wind speeds, slope, elevation, etc.).The altitude of our research area ranged from 3600 to 4400 m, which is much higher than other study sites; thus, the temperature is lower than that in other study areas, which caused the tree growth in a relative lower elevation area to show a significant positive relationship with the temperature (DPL, DPH).The increase in the elevation and hydrothermal combination gradually changed and caused the change in the ring widthclimate relationship.Wind speeds are higher at high-elevation sites in the STP [24], and the wind speed data acquired from automatic weather stations showed that the wind speeds at the DPH are approximately twice as high as those at the DPL, but precipitation at the two sites is roughly equal (Figure 7).High wind speeds and increasing evapotranspiration can cause drought stress, which occurred in the dry season.In addition, the slope in our sampling site at high elevations is greater than that in middle and low elevations; thus, a lot of precipitation is lost in the form of runoff.Therefore, precipitation and the SPEI became the main climatic factors to influence tree growth in high-elevation areas (DPH, DPF, ZWH).
The second possible reason concerns the biotic factors (the depth of root, forest type, tree age, and sensitivity, etc.).P. likiangensis is a shallow-rooted plant which is more susceptible to frost damage [71].Thus, in the DPL and DPM, the ring growth of P. likiangensis is tightly correlated with temperature.Different forest types, such as open, closed, and mixed forests, significantly influence how trees respond to climatic variables.All five sites studied in this research are categorized as open forests, which have been observed to be more sensitive to temperature fluctuations compared to closed forests [72].Unlike the DPL and DPM, the DPH is a mixed forest, where interactions between various tree species can modify responses to climate change [73].This also explains one possible reason for the variation in the relationship between the TRI and climate along different altitudes.As to the age of trees, we can see that at the DPL, the age of these trees is relatively younger compared to the age of trees at the DPM and DPH, and there is no decay inside trees collected in the DPL.At the DPM and DPH, the interior of the older trees had decayed.As a result, the samples could not reflect the true age of the whole trees in the DPM and DPH.However, the age of the tree can be deduced by its trunk diameter and height [74].In the DPH, both the trunk diameter and height of the trees are significantly larger than those in the DPL.So, there is no doubt that trees growing at higher altitudes are older than trees growing at lower altitudes (Figure 3).The age of trees plays a crucial role in their response to climate change.For example, the basal area increments (BAIs) of trees, which are influenced by tree size and age, have a negative correlation with the tree-ring mean sensitivity [75].Many other studies have also verified that the age of trees can affect the M.S. of radial growth, with the radial growth of older trees being more sensitive to climate change than younger trees [76][77][78].In addition, large trees have different physiological characteristics compared to intermediate trees [79].So, the age of trees may play a significant role in the changes in the ring width-climate relationship with altitude.
Previous studies indicated that M.S. is a measure of year-to-year growth variability, which is generally considered to reflect the growth sensitivity to high-frequency climate variability; and S.D. reflects the multi-decadal growth variability in a chronology related to low-frequency climate variability [10,80,81].Therefore, these two indicators can be seen as an aspect of growth variability.The M.S., S.D., SNR, and EPS of chronology were key indicators to test the reliability of the chronology and strength of the climate signal contained in TRW (Table 2).In this study, we found that along the studied elevation gradient, these parameters all displayed a wave-like tendency of increasing and then decreasing values.A similar trend was reported in other studies conducted in the STP and other places in China [14,22,23,34,82,83].For instance, in Naqu, southern TP, He et al. [23] sampled increment cores of Juniperus tibetica at four elevations along altitudinal transects from 4000 to 4500 m a.s.l.; the results of this study showed that the values of SNR, EPS, and M.S. were the lowest in the high juniper belt and the highest in the mid-high juniper belt.Panthi et al. [34] developed tree-ring width chronologies of Abies georgei along elevation

Figure 1 .
Figure 1.Location of study area (a) and sampling sites (b), and the landscape of study area (c).The blue circle represents Daocheng meteorological station, the red triangle represents the study area, the five-pointed stars represent the location of the five sampling sites, and the red arrows indicate our sampling route (from 3600 to 4400 m a.s.l.).DPH/DPM/DPL represent the upper distribution limit/the middle range/ the lower distribution limit of Picea likiangensis.DPF/ZWH represent the sampling site of Sabina saltuaria and Abies squamata.

Figure 1 .
Figure 1.Location of study area (a) and sampling sites (b), and the landscape of study area (c).The blue circle represents Daocheng meteorological station, the red triangle represents the study area, the five-pointed stars represent the location of the five sampling sites, and the red arrows indicate our sampling route (from 3600 to 4400 m a.s.l.).DPH/DPM/DPL represent the upper distribution limit/the middle range/the lower distribution limit of Picea likiangensis.DPF/ZWH represent the sampling site of Sabina saltuaria and Abies squamata.

Figure 1 .
Figure 1.Location of study area (a) and sampling sites (b), and the landscape of study area (c).The blue circle represents Daocheng meteorological station, the red triangle represents the study area, the five-pointed stars represent the location of the five sampling sites, and the red arrows indicate our sampling route (from 3600 to 4400 m a.s.l.).DPH/DPM/DPL represent the upper distribution limit/the middle range/ the lower distribution limit of Picea likiangensis.DPF/ZWH represent the sampling site of Sabina saltuaria and Abies squamata.

Figure 2 .
Figure 2. Monthly variation in total precipitation (blue bars), mean maximum temperature (line with green triangles), mean temperature (line with purple squares), and mean minimum temperature (line with red circles) for Daocheng meteorological station of southwestern Sichuan, China.

Figure 2 .
Figure 2. Monthly variation in total precipitation (blue bars), mean maximum temperature (line with green triangles), mean temperature (line with purple squares), and mean minimum temperature (line with red circles) for Daocheng meteorological station of southwestern Sichuan, China.

Figure 3 .
Figure 3. Variations in annual mean temperatures (a) and annual precipitation (b) from 1959 to 2020 at Daocheng station.

Figure 3 .
Figure 3. Variations in annual mean temperatures (a) and annual precipitation (b) from 1959 to 2020 at Daocheng station.

Figure 4 .
Figure 4. Age distribution of Picea likiangensis along elevations and the ring width.The sub-graphs (a-c) are represented as DPL, DPM and DPH, respectively.In subfigure (d), green/red/purple lines represent the age curves at DPH/DPM/DPL, respectively.

Figure 4 .
Figure 4. Age distribution of Picea likiangensis along elevations and the ring width.The sub-graphs (a-c) are represented as DPH, DPM and DPL, respectively.In subfigure (d), green/red/purple lines represent the age curves at DPH/DPM/DPL, respectively.

Figure 5 .
Figure 5.Standard ring width index chronologies along elevation from southeastern Tibetan Plateau.The sub-graphs (a-e) are represented as DPL, DPM, DPH, ZWH, and DPF, respectively."Number" represents the sample depth (gray shading), TRI stands for tree-ring index (blue bold line), Rbar represents the mean series inter-correlations (black line), and EPS stands for expressed population signal (red line).The red dashed line indicates EPS = 0.85.

Figure 5 .
Figure 5.Standard ring width index chronologies along elevation from southeastern Tibetan Plateau.The sub-graphs (a-e) are represented as DPL, DPM, DPH, ZWH, and DPF, respectively."Number" represents the sample depth (gray shading), TRI stands for tree-ring index (blue bold line), Rbar represents the mean series inter-correlations (black line), and EPS stands for expressed population signal (red line).The red dashed line indicates EPS = 0.85.

Figure 6 .
Figure 6.Results of Pearson correlation analysis between tree-ring chronologies and climatic factors in DPL (a), DPM (b), DPL (c), ZWH (d), and DPF (e).CC represents correlation coefficient.Tmean/Tmax/Tmin, P, and SPEI represent the monthly mean/maximum/minimum temperature, precipitation, and standardized precipitation evapotranspiration index, respectively.The green dashed line and the red dashed line represent 95% and 99% confidence levels.PMay represents the previous May and PN-A represents the previous November to current Augst.PO-Jun represents the previous October to current June.PS-Jan represents the previous September to current January.PO-F represents the previous October to current February.

Figure 6 .
Figure 6.Results of Pearson correlation analysis between tree-ring chronologies and climatic factors in DPL (a), DPM (b), DPH (c), ZWH (d), and DPF (e).CC represents correlation coefficient.T mean /T max /T min , P, and SPEI represent the monthly mean/maximum/minimum temperature, precipitation, and standardized precipitation evapotranspiration index, respectively.The green dashed line and the red dashed line represent 95% and 99% confidence levels.P May represents the previous May and PN-A represents the previous November to current Augst.PO-Jun represents the previous October to current June.PS-Jan represents the previous September to current January.PO-F represents the previous October to current February.

Figure 7 .
Figure 7. Relative humidity (RH), precipitation, and wind speed data acquired from automatic weather stations (HOBO U30, USA) in DPH and DPL from 26 March 2023 to 26 August 2023.24-Apr represents the date 24 April 2023.

Figure 7 .
Figure 7. Relative humidity (RH), precipitation, and wind speed data acquired from automatic weather stations (HOBO U30, USA) in DPH and DPL from 26 March 2023 to 26 August 2023.24-Apr represents the date 24 April 2023.

Table 2 .
Statistics of standardized tree-ring width chronologies.

Table 2 .
Statistics of standardized tree-ring width chronologies.
T/C: core/tree in the common interval, M.S.: mean sensitivity, SNR: signal-to-noise ratio, EPS: expressed population signal, WTR: mean correlations within-trees, BTR: mean correlations between trees, and Rbar: mean series inter-correlations.