How tall is forest

Tree-rings were cross-dated visually using the skeleton plot method 34 , and the ring widths were measured with an accuracy of 0. Some cores were excluded because they could not be well measured or cross-dated. We then constructed a chronology for each plot, using the remaining 23—32 cores Table 1. The growth trends decreasing ring-width with increased age were removed from raw ring-width data with exponential curves, which was necessary for examining the high-frequency growth variation associated with climate changes A residual chronology that maximized the climatic signal was then created for each forest type using the ARSTAN software This suggests that the qualities of the chronologies were good enough for further study Monthly climate data between and of the Erdao meteorological station, at the foot of Mt.

Our plots differed greatly in altitude and thus, for a better analysis of the growth-climate relationship, we estimated the historical monthly climate for each plot.

For altitudinal changing rates of monthly mean temperature and total precipitation see Supplementary Table S1 , we used the model of Wang et al. The model was developed based on 48 climate stations in the Mt. For each plot, the climate variables in each month by each year to were estimated based on the altitudinal changing rates of temperature and precipitation in Supplementary Table S1. Then the RWI of the plot was related to these monthly climate variables to examine the growth response to climate change see Supplementary Table S2.

We also calculated the multi-decades means of monthly temperature and precipitation for each plot, which were then used to calculate climate indices such as actual evapotranspiration AET , mean annual temperature and precipitation, mean temperature for the coldest month, etc.

These multi-decades mean climate indices were used to depict the climate condition of each plot, in order to evaluate the relative effect of climate gradient vs. We used MS and SD of the chronologies as indicators of temporal growth variability of radial growth. For growth-climate relationship, we related RWI of each plot to monthly temperature and precipitation from June of the previous year through September of the current year Supplementary Table S2.

S1 , using the correlations between RWI and monthly climate variables of the 15 chronologies. As shown in Supplementary Fig. Thus the PCA results confirmed that the correlations we selected were major components of altitudinal change of growth response to climate change, and are well suited for our purpose. While there were still some other significant correlations in Supplementary Table S2 , they did not reveal clear altitudinal patterns or occurred in only one or two plots.

These correlations may be caused by local abiotic and biotic factors specific to the plots, which were not our focus and not further analyzed. To examine the potential drivers for altitudinal changes of growth variability MS, SD and RWI-climate correlations, we used general linear models GLM to explain these variables with explanatory terms as follows.

We used AET to depict the climate condition of the plots, because AET reflects the simultaneous availability of energy and water and is widely used as proxy of climate productivity of vegetation 26 , Other climate variables, such as mean annual temperature and precipitation, were highly correlated with AET and thus not used to avoid collinearity. Here we used the maximum tree height of the plot, which is a commonly used surrogate e. We also used the mean height for the trees that tree-rings were sampled to repeat the statistical analyses, and the results were similar as using forest height.

This is not surprising because we cored dominant canopy trees. Previous studies have found that RWI-climate relationship differed between large- and small-DBH trees 6 , 21 , suggesting the potential role of tree size on growth response to climate change. In the preliminary statistical analyses, we have included interactions in GLMs.

However, the interaction terms were not significant or showed only weak effects, and thus were excluded from the final analyses. To identify the most important factors affecting growth variability and RWI-climate relationships, we first used the model selection approach based on Bayesian Information Criterion BIC to obtain the most parsimonious model, which is more conservative in retaining variables than commonly used Akaike Information Criterion AIC The statistics for chronology of each plot were reported in Table 1.

For the two indicators of temporal growth variability, MS ranged between 0. Changes of two chronology statistics, i. There were six correlations that showed significant altitudinal patterns i. Climate in the previous summer also showed clear altitudinal patterns. Changes of correlations between ring-width indices and some monthly climate variables along altitudinal gradient.

Only forest height was significantly correlated with the AET gradient across altitudes. The six RWI-climate correlations were more closely related with forest type and climate R 2 between 0. Forest type was retained in most models and had the highest importance value 0. AET was retained in all the models, and its importance 0. Despite lower importance 0. On the other hand, maximum volume was excluded from all the models.

Previous tree-ring studies along geographic gradients generally focus on the change of RWI-climate relationship, instead of the temporal growth variability during climate change. Our study sampled 15 plots across a large altitudinal gradient from temperate to timberline forests. We showed that growth variability MS and SD increased at higher altitude on humid mountains, even when different altitudinal zones were included Fig.

Previous studies have noticed that MS and SD changes with altitude. In humid regions MS and SD were generally found to be higher with increasing altitudes 14 , 16 , 48 , while the converse altitudinal pattern was commonly observed in arid regions 10 , 27 , This difference is generally explained by the fact that: in arid regions, the water deficit at low altitudes is critical in limiting radial growth, but the stress is alleviated at higher altitudes and thus growth variability is lower 14 , Thus, the two contrasting altitudinal patterns are actually consistent, in that MS and SD are higher under harsher climate In addition to changes along altitude, other studies also found that MS and SD differed markedly among forest types species under similar climate e.

For instance, Liang et al. These results suggest that species identity may be another important factor affecting MS and SD in addition to environmental condition. With these previous results, we had expected that climate and forest types would be the strongest predictors for MS and SD. However, only forest height was retained in the final models Table 4.

This result does not mean that climate and forest type have neglectable influence on MS and SD. Nevertheless, our results do suggest that forest height is a very important predictor for the temporal variability of radial growth.

Forest tree height has seldom been used to explain MS and SD before. However, in a few tree-ring studies that have documented forest height or tree size , we also found evidence that MS SD decreased with taller forest height For instance, in a study that MS and SD increased with higher altitude on a humid mountain, forest height decreased sharply with increasing altitude Similarly, on an arid mountain in northwest China, MS and SD decreased while tree height increased with higher altitude In another study along a successional series, MS and SD decreased from early- to late-successional forests, accompanied by an increase in forest height towards late-successional forests Nevertheless, the potential role of forest height was not explored in these studies.

In all these cases MS and SD were negatively related to forest height or tree size, similar to our results Table 3.

As for why growth variability MS and SD is negatively related to forest height, we hypothesis that this is because radial growth will show higher variability when productivity is lower. Forest height is well known to be higher in more productive sites e. Both forest productivity and height decrease with higher altitude on humid mountains but increased on mountains in arid regions.

Consequently, MS and SD increased with higher altitude on humid mountains e. This hypothesis may also explain why forest height is a better predictor in our study. Forest productivity is affected by a number of factors in addition to climate, e. Meanwhile, forest height is also affected by both these local factors and climate, and thus may a better indicator of productivity than climate indices.

Along successional series, forest productivity generally increases from early- to late-successional stages e. According to our hypothesis, MS and SD should be lower in later-successional forests, and thus lead to a negative relationship of MS and SD with forest height.

This prediction is consistent with the findings in Liang et al. Meanwhile, larger-DBH trees i. Consequently, smaller-DBH trees are expected to have higher MS, which has been reported in previous studies 21 , 51 , Thus, our hypothesis seems to provide an explanation for different situations altitudinal gradients in humid and arid regions, successional series and different tree sizes.

However, the above-mentioned hypothesis and still needs further tests. Geographic patterns of MS and SD seem to be related to a number of abiotic and biotic factors, such as temperature gradient, water availability, tree size, age and species identity etc. Height is a key dimension of tree size, and is closely related to not only DBH and tree age but also productivity. Thus in theory height ought to have a critical in affecting the growth response to climate change. We suggest that more studies are needed to test the relative influence of tree height, age, species identity and climate, for a better understanding of temporal growth variability under climate change.

Many studies have reported that RWI-climate relationship changes along altitudinal gradients. Summer temperature was generally found to be the limiting climate factor of tree growth at high altitudes, while water-availability was the limiting factor at low elevations 9 , 10 , 11 , 14 , Our results on the altitudinal change of RWI-climate correlations are generally consistent with these previous findings. At low altitudes, RWI was positively related to current June precipitation while negatively related to current June temperature, a typical indication that water availability is limiting for radial growth e.

This limiting effect disappeared towards high altitude, which is a commonly-reported pattern because of increased precipitation and decreased temperature with higher altitude 9 , 10 , Influence of previous summer climate on current year growth i. Interestingly, previous and current summer climate showed not only contrast effect on growth but also converse altitudinal pattern Fig.

Similar contrast effects of current and previous summer climate on radial growth have also been found in previous studies in humid regions e. Our results support the idea that these altitudinal patterns can not be explained by climate gradient alone. Forest type had the highest importance in each model Table 4. Thus, in contrast to previous studies that have included only one forest zone or species, we conclude that forest type is the most important factor leading to different RWI-climate relationship, while altitudinal climate gradient plays a second role.

Recent studies along latitudinal have also suggested that the geographic change of RWI-climate relationship is shaped by climate gradient and forest type together 3 , 4 , 9 , Local scale studies have noticed the effects of tree size age on RWI-climate relationship. For instance, older trees were found to be more sensitive to climate change than younger ones 22 , 61 , 62 , and large-DBH trees were more sensitive to drought than small trees 6 , Some authors also showed that the differences in RWI-climate relationships between DBH classes were greater than those between species These evidences suggest that tree size or height may play a role in affecting RWI-climate relationship.

As for the effect of stand density, forest thinning treatments have found that low density stands were less sensitive to droughts than high-density ones, because lower competition increases the resource availability of individual trees 6 , Based on these evidences, we had expected that forest height, maximum volume and stem density might play a stronger role in explaining RWI-climate relationship.

However, our results showed that the importance of height, volume and density were low Table 4. Studies that found clear effect of tree size age and density were generally conducted by comparing trees of a same species under a similar climate.

Our results suggest that height and density are weaker modulators of RWI-climate relationship, when compared with the great gradients of climate and forest type. It should be noted that forest height and density entered most of the final models for the RWI-climate correlations, suggesting they still played a necessary role in addition to forest type and climate. Another possible reason for the weak effect of height and density found here may because our study area is a humid cold region without water deficit 25 , Existing evidences on the effect of tree size age and density were mostly found for the growth sensitivity to drought 6 , 21 , Thus, we suggest that forest height and density may reveal a much stronger effect in arid regions, which need to be tested carefully in the future.

In summary, our study showed that both growth variability and RWI-climate relationships varied regularly across different forest types along an altitudinal gradient in northeast China.

We also examined the relative importance of climate, forest type, height and density in shaping these altitudinal patterns probably the first one along altitudinal gradients. We showed that forest height is the most important predictor for MS and SD, while forest type and climate were more important for RWI-climate relationship with height and density still played a necessary role. The importance of forest type and height highlight the necessity for future studies to test the effect of forest structure species composition, height and density, etc.

If these effects are verified across latitudinal altitudinal gradients, then forest structure should be considered when predicting global forest growth in response to future climate change. Climate Change Impacts, Adaptation, and Vulnerability. Cambridge University Press, Lara, A. Article Google Scholar. Huang, J. Radial growth response of four dominant boreal tree species to climate along a latitudinal gradient in the eastern Canadian boreal forest. Global Change Biol. Babst, F.

Global Ecology and Biogeography 22 , — EO Explorer. At the time of publication, it represented the best available science. View this area in EO Explorer. View more Images of the Day:. You might also be interested in view all.

Subscribe to our newsletters. Two main opposing forces affect a tree's height; one pushes it upward while the other holds it down. By analyzing the interplay between these forces, a team of biologists led by George Koch of Northern Arizona University calculated the theoretical maximum tree height, or the point at which opposing forces balance out and a tree stops growing.

This point lies somewhere between and feet and m. On the one hand, the researchers found, trees in forests "desire" to grow as tall as possible to overtake neighboring trees and reach stronger sunlight.