Drivers of tree demographic trade-offs in a temperate forest

Jianghuan Qin, Chunyu Fan, Yan Geng, Chunyu Zhang, Xiuhai Zhao, Lushuang Gao

Research Center of Forest Management Engineering of State Forestry and Grassland Administration, Beijing Forestry University, Beijing, 100083, China

Keywords:Tree survival Density dependence Growth-survival trade-offs Life-history strategy Temperate forest

ABSTRACT

1. Introduction

Forest dynamics and community assembly are often explained by inter- and intraspecific variations in demographic performance (Rees et al., 2001; Russo et al., 2010; Gadow and Kotze, 2014; Adler et al.,2018; Keram et al., 2021). Fundamentally, demographic rates, such as species' growth, and mortality along environmental resource gradients can ultimately influence species diversity and distribution. Tree growth and survival are considered to be two of the most important vital rates of tree demography, and they are the results of the combined effects of biotic and abiotic factors.The interspecific demographic trade-off exists between a species' ability to grow quickly in ‘‘favorable conditions’’(when resource availability is high and other stresses are absent) vs. its ability to avoid mortality in ‘‘unfavorable conditions’’ (when resource availability is low or other stresses are present) (Wright et al., 2010).Several approaches have been proposed to model the growth-survival relationship, varying in the data sources and model flexibility (Russo et al., 2021). However, the mechanisms' effect on the growth-survival relationship and how they vary across life stages are rarely studied due to a lack of long-term data (Wunder et al., 2008). Tree growth and survival are key elements of forest dynamics,and thus are of great concern to forest managers,yet our understanding of the mechanisms leading to the growth-survival trade-off pattern remains limited.

Drivers of the growth and survival trade-off have been a central topic for understanding the structure and function of temperate forests. A growing number of studies involving tropical and temperate forests support the idea that neighborhood interactions such as conspecific negative density dependence (hereafter “CNDD”) and environmental filtering are key factors in structuring tree species composition and diversity turnover in forests (Comita et al., 2014; Ramage and Mangana,2017;Zhu et al.,2018;Song et al.,2021).CNDD,which was proposed by Janzen and Connell(Janzen,1970;Connell,1971),generally shows that the closer spatial distance among individuals will produce more intense competition for resources. At the same time, spatial aggregation of the same species is more likely to have disease transmission,and to be preyed or eaten, which is unfavorable to their offspring (seed or seedling)growth. It also indicates that the survival and growth of individuals or propagules decline with increasing density of conspecific neighbors(Comita et al., 2014). While CNDD has been widely observed to affect tree growth and survival across temperate and tropical forests, the strength of CNDD and resource filtering varies substantially among species with different life-history strategies within a community (Comita et al.,2010; Johnson et al.,2012).

There is clear evidence of trade-offs for a species between tending to grow quickly under high light conditions and to survive in deep shade(Comita et al.,2009).Shade-tolerant tree species with slow growing rates tend to have higher wood density and superior defenses compared to fast-growing, and light-demanding species (Wright et al., 2010). Additionally, the negative effects of conspecific neighbors were stronger for light-demanding species than shade-tolerant species(Comita et al.,2009;Qin et al.,2020).Moreover,the varying strength of conspecific neighbor effects has also been linked to tree species mycorrhizal type (Crawford et al., 2019). The type of mycorrhizae associated with plant hosts is an important root functional trait that can affect plant nutrient uptake,and will likely influence species' CNDD strength through microbiomes(Bennett et al.,2017).AM(arbuscular mycorrhizal)trees mainly require soil inorganic nutrients, while EM (ectomycorrhizal) trees exhibit a competitive advantage in accessing and absorbing organic nutrients(Liu et al.,2018).Similarly,in a previous analysis,Chen et al.(2019)showed that mycorrhizal type mediated tree neighborhood interactions at the community level in a subtropical forest,while Jiang et al.(2020)and Jia et al.(2020)found that tree mycorrhizal association may determine the strength of CNDD's effects on temperate forests.

Above- and below-ground traits and interactions may strongly influence the strength of CNDD and environmental filtering,according to the performance mentioned above. Attempts to identify how the variations in CNDD impact growth and survival have produced conflicting results.And it may reflect the combined effect of CNDD with species trait characteristics, and abiotic environmental filtering on tree growth and survival rates(Zhu et al.,2015b,2017;Comita,2017).In the present study,we sought to further assess the effects of biotic and abiotic factors on the growth and survival relationships in a temperate forest in northeastern China to quantify the strength of CNDD and its effects on the growth-survival trade-offs at multiple life stages, from saplings to adult trees, for 24 temperate tree species. We mainly focus on the roles of CNDD in driving plant growth and survival trade-offs that are not well examined at different tree life-history strategies and multiple life stages.We made the following specific hypotheses:(1)The drivers of individual survival change with life stage and life-history strategy; and (2) CNDD drives the trade-offs of species' growth-survival with environmental stress. We also examine biotic and abiotic effects on tree individual survival and how the effects of CNDD on the growth-survival trade-off vary across the life-history strategy among tree species in the community.

2. Materials and methods

2.1. Study site

The plot represents a mixed broad-leaved Korean pine (Pinus koraiensis) community which covers a 42-ha (840 m × 500 m) temperate forest and is located in Jilin Province, northeastern China (43°58′N,127°45′E).The average annual temperature and average annual rainfall in the study area was 3.8°C, and 695.9 mm, respectively. The average temperature was-18.6°C during the coldest days in January,and 21.7°C during the hottest days in July. The plot has an elevation ranging between 459 and 517 m a.s.l. and it was considered to represent the typical and most common forest type of that region, and characterized with a temperate continental monsoon climate. The five dominant species are Juglans mandshurica, Acer mono, Tilia amurense, Fraxinus mandshurica and Pinus koraiensis. All free-standing woody plants with diameter at breast height(DBH)≥1 cm were mapped and identified to species.And DBH,tree heights and crown widths of all individual plants were measured in the plot during the summers (July and August) of 2010, 2015 and 2020, when the first, second and third full assessments respectively took place. The species information can be found in Table S1.

2.2. Data preparation

To improve the accuracy of our calculation,we only selected species that have more than 100 individuals observations in the 2010 census,for further analysis. There are 24 species (excluding liana and shrubs) in total whose data meet the requirement.In the present study,the data of the two censuses (2010–2015, 2015–2020) were used to determine the mortality rates for all species and relative growth rates of the DBH for each tree in the research site.

Relative growth rates(RGR):

where DBHt1and DBHt2represent the DBH values in the 2010–2015 census and 2015–2020 census,respectively;Nt1is the initial individuals in the 2010–2015 census; and Nt2indicates the survivors in 2015–2020 census.

For each species, we used the 90-percentile relative growth rate(RGR90)as a proxy for growth rates under favorable conditions(Wright et al., 2010). The mortality rates under unfavorable conditions were calculated for the 25% of individuals of each species with the smallest RGR in the previous census interval(MR25) (Wright et al.,2010).

We define neighborhood competition(A)as the sum of the tree basal areas (BA) divided by the distances to each tree from every focal tree within a 20-m radius(Bai et al.,2012;Yan et al.,2015;Qin et al.,2020;Yao et al., 2020):

where i represents a conspecific(Acon)or heterospecific(Ahet)individual within the area of a 20-m radius around a focal tree. The biotic neighborhoods were derived from the basal areas recorded in 2015.

All of the topographical variables (elevation, convexity, slope and aspect)were calculated and soil samples were collected from each 20 m×20 m subplot.The nine soil nutrient indexes including pH,soil water,OM,TN,TP,TK,AN,AP,AK(see Table 1 for abbreviation terms)of each sample were measured and utilized in our analyses as described previously in our methods(see Qin et al.,2021 for more details).Next,we used a principal component analysis (PCA) to reduce model complexity and avoid collinearity among the variables.Finally,we used PC1 to represent the topography, soil elements and leaf area index.

2.3. Data analysis

We used logistic generalized linear mixed-effects models (GLMMs;Bolker et al.,2009)with binomial errors(i.e.tree alive coded:1,or dead:0)to examine the influence of neighborhood drivers on the probability of survival(Qin et al.,2021).Since the initial size of a tree can significantly affect tree growth and survival (Comita et al., 2009; Wang et al., 2012;Piao et al., 2013), we modeled survival over the most recent 5-year census interval (2015–2020) to examine the influence of neighborhood densities and RGR(2010–2015),as well as their interactive effects.

Table 1 Abbreviations and their associated terms used throughout this paper.

Species shade tolerance and mycorrhizal type were grouped according to Wang et al. (2009) and Qin et al. (2021). We divided the tree species into two groups (shade-tolerant species and shade-intolerant species) due to shade tolerance is hard to classify (Niinemets and Valladares, 2006), especially mid-tolerant species. For each guild (life--history strategy), we separately modeled the probability of individual survival of the second census interval(2015–2020)as a function of RGR,conspecific neighbor density (Acon), heterospecific neighbor density(Ahet), the interactions between the conspecific effect and RGR,and the interactions between heterospecific effect and RGR. To indicate the importance of abiotic factors,we also included PC1 as a fixed effect in the model. All continuous independent variables were standardized before entering the model.In addition,we included species identity as a random effect,and we assigned the quadrat where the individual was located as a random effect because individuals close to each other may have similar survival and growth probabilities(i.e.spatial autocorrelation)(Chi et al.,2015;Zhu et al.,2015b).

For tree species, we performed the linear regression analyses to evaluate the relationships of RGR-MR and RGR90-MR25,respectively. In addition, to evaluate the growth and survival trade-offs, we performed principal components analyses (PCA) for the species relationship between RGR90–MR25and RGR-MR (Wright et al., 2010). In subsequent analyses, species positions on the growth-mortality trade-off equaled their factor score on the first two principal components of the RGR90-MR25relationship and RGR-MR relationship, respectively. Besides, we respectively compared the associations between species conspecific effect (CNDD) and the two factor scores to determine the strength of density dependence in driving community dynamics.

The GLMMs were executed using “glmer” function in the “l(fā)me4”package(Bates et al.,2015)and we created graphs using the functions of the“ggplot2”and“ggbiplot”package.All analyses were conducted using R version 4.1.2(R Core Team,2021).

3. Results

We got data of 38,876 individuals in 2020 and 46,631 individuals in 2015. A total of 24 canopy species differing in light requirement (15 shade-intolerant species vs. 9 shade-tolerant species), mycorrhizal type(12 arbuscular mycorrhizal species vs.12 ectomycorrhizal species)were collected.At the same time,all of the individuals were classified as three life-stages (saplings, juveniles and trees), and the demographic rates(such as RGR and MR of each stage)were shown in Table S1.

3.1. Conspecific effect on survival relationships of individuals

Overall,RGR exerted a significant positive effect on the probability of survival except for shade-intolerant (Not_shade) species and juveniles(Fig.1a).Significant negative effects of conspecific neighbors on survival were detected for the community level, and early life stages with both saplings and juveniles showing a lower probability of survival at higher conspecific densities (Fig. 1b). Shade-intolerant (Not_shade) species suffered more negative conspecific effects than shade-tolerant species,and EM trees also showed more negative effects than AM trees(Fig.1b).In contrast, heterospecific tree neighbor density tended to be nonsignificantly for all groups except juveniles (Fig. 1c). Environment variables showed non-significant negative effects on juveniles and trees(Fig. 1d). A lack of statistical significance relationships was shown between abiotic environmental factors and later life-stages (Fig. 1d). We detected all groups with negative interaction between RGR and conspecific neighborhood density for survival; however, the interactions were non-significant in the three life-stages (saplings, juveniles and trees;Fig. 1e). This illustrated significant negative effects on overall, species shade tolerance and mycorrhizal type (Fig. 1e). As for the RGR's interactions with the heterospecific neighborhood density, we only found significant negative relationships in shade-tolerant species and AM trees(Fig.1f).

3.2. Growth-mortality trade-offs among species

The growth-mortality trade-offs were found to be significant for the relationship between MR and RGR (Fig. 2b). The coefficient of determination for the relationship between the RGR90and MR25was only 0.0094(p=0.65),as detailed in Fig.2a.The trade-offs between the RGR and MR were shown to be significantly stronger than those between the MR25and RGR90,and the differences in the correlation coefficients were found to be significant(Fig.2b; R2=0.53;p ＜0.001).

The first score (PC1) of PCA represented 54.8% and the RGR90and MR25had an inverse relationship with PC1 and PC2 represented 48.2%which showed a consistent relationship of RGR90-MR25trade-offs(Fig. 3a). However, both RGR and MR were positive with PC1 for a high explained 86.3% and PC2 represented only 13.7% and showed a converse trend(Fig.3b).

3.3. Conspecific effect on species growth-survival trade-offs

The CNDD showed a consistently positive linear relationship with RGR90-MR25factor score 2(PC2)and RGR-MR factor score(PC1),which signified that CNDD had a strong effect on the growth-survival trade-off at community-level (Fig. 4b and c). However, CNDD showed a nonsignificant relationship between RGR90-MR25factor score (PC1) and RGR-MR factor score 2 (PC2) (Fig. 4a and d), and the PCA results of RGR90-MR25and RGR-MR may result in this difference(Fig.3).

4. Discussion

4.1. Drivers of tree demographic trade-offs

Fig.1. Odds ratio(95%confidence)of(a)individual RGR(species relative growth rate),(b)conspecific neighbor density,(c)heterospecific neighbor density,(d)PC1(first axis of PCA for abiotic factors), (e) interactions between RGR and conspecific, and (f) heterospecific neighbor density on individual survival at the different species life-history strategy and sapling, juvenile and adult stages in the forest. Survival was modeled separately for each groups using GLMMs (see Methods). Solid circles indicate significant effects(p ＜0.05).Overall:community level;Shade:species with shade-tolerant;Shade_in:species with shade-intolerant;AM:species with arbuscular mycorrhizal; EM: species with ectomycorrhizal; saplings: DBH ≤10 cm; Juveniles: 10 cm ＜DBH ≤20 cm; Trees: DBH ＞20 cm.

Fig. 2. Growth and mortality trade-offs expressed as(a) the 90th percentile relative growth rate (RGR90)vs. the mortality rates of the slowest growing 25% of the individuals (MR25), and (b) the mean RGR and MR for the 24 tree species; type-AM and EM mean species mycorrhizal type.The demographic rates were calculated with all individuals of each species. The legend shade abbreviation of not_shade is shadeintolerant, and shade represents shade-tolerant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Our results showed that CNDD and environmental filtering across the tree characteristic and life-stage gave the key step in shaping community composition and maintaining diversity.As anticipated, we found significant negative conspecific effects from the overall level, which was consistent with the results of related research in tropical and temperate forests (Bai et al., 2012; Zhu et al., 2018; Qin et al., 2020; Yao et al.,2020). CNDD may be the vital mechanism for the maintenance of tree species diversity even though other mechanisms like resource niche partitioning, and dispersal limitation are likely to be occurring simultaneously because CNDD will likely not occur in isolation in temperate forests (Barry and Schnitzer, 2021). Furthermore, the CNDD showed a direct effect on both RGR-MR and RGR90-MR25trade-offs,indicating that density dependence may drive community dynamics in temperate forests. Another possible explanation was that the fastest and slowest growth rate did not indicate that the individuals of the species were living in either favorable or unfavorable environmental conditions. It is nebulous to examine the growth and mortality trade-offs for all individuals in comprehensive environmental conditions. For example, a species living in a site that is under stress may show low mortality rates by adopting specific ways to utilize the limited available resources(Fan et al.,2017). This suggests that the inherent differences among the species could be of major importance for the tree growth and mortality trade-offs(Rüger et al.,2012).

Fig.3. Principal components analyses for(a)RGR90 and MR25,and(b)RGR and MR for abundance of more than 100 of 24 tree species in 42 ha forest dynamics plot.

Fig. 4. Relationship between conspecific effect and the factor scores of growth-mortality trade-offs. The legend abbreviations are the same as those in Fig. 3.

We observed that species demographic trade-offs also varied significantly in response to abiotic variables(Fig.1).This result indicated that individuals with different life-history strategies varied widely to resource partitioning processes. For individuals in the shaded understory, which includes most of the seedlings and saplings,this trade-off is likely driven predominantly by light availability. However, the growth of adult tree species is likely not as strongly light-limited,as they are located closer to or within the canopy (Wright et al., 2010). Species growth-mortality trade-offs may be driven by both CNDD and environmental filtering not only at community level but also population to current species coexistence through differential responses to abiotic variables in the forest plot (Fig.4).

The influences of biotic and abiotic environmental variables on both growth and survival further alter this developmental community. We hope that understanding how the features of a community development determine population dynamics may guide future research regarding specific mechanisms which will unlock this “black box”, such as pathogens and resources and their interactions.

4.2. Relationship of growth-survival trade-offs

The relationship we found between species RGR and CNDD was also along with previous study(Zhu et al.,2018).For individuals,our results showed that the RGR had a positive effect on tree survival, which was consistent with previous results(Russo et al.,2007).Higher growth rates of individual trees are likely associated with increased access to resources; for example, larger trees generally have more access to light(Fien et al., 2019) and canopy trees may have significantly dispersal abilities to avoid stronger CNDD(Barry and Schnitzer,2021).

Our results showed a positive relationship between tree species'growth and mortality rates in this temperate forest. However, it was determined that the trade-offs of the RGR90and MR25were nonsignificant, which revealed that the variations in resource availability among the species may not have been significant enough to detect.Iida et al.(2014)found that positive correlations between RGR and MR were present over the whole range of stem diameters, but they were only significant at small stem diameters.Wright et al.(2010)found that there was an interspecific trade-off between rapid growth under favorable conditions and low mortality under unfavorable conditions for small trees,but not for large trees,in tropical forests on Barro Colorado Island(BCI), Panama. The results of our study are similar to those of recent studies performed in temperate and subtropical forests of China (Wang et al.,2012;Wu et al.,2017),which found that tree mortality increased with growth rate at the community level.

Consistent with the results of the previous study(Zhu et al.,2018),we found a significant negative interaction between RGR and CNDD in individuals' survival. This is somewhat surprising, given that CNDD is generally more strongly linked to life-history strategy. For example,fast-growing shade-intolerant species may be more susceptible to pathogens which attack large trees and increase mortality risk due to lower wood density (Zhu et al., 2018). Interspecific growth-survival trade-off also shows the consistent result that CNDD may exert a crucial impact on population dynamics with species characteristics.

4.3. Demographic trade-offs and life-history strategies

Understanding the factors affecting dynamic and complex processes,such as growth and mortality, is necessary to successfully manage uneven-aged, mixed-species forest systems for continued resilience and productivity.And the results of our study demonstrate that such variation among species in the strength of conspecific neighbor effects is not likely random, but rather it is driven at least in part by species' life-history strategy rather than life stage (Fig. 1e), as Zhu et al. (2018) revealed.Consistently,there are increasing recognitions that the strength of CNDD varies widely among communities; this is most likely the result of, for example,differences in allocation related to shade tolerance(Valladares and Niinemets, 2008; Brown et al., 2020) and mycorrhizal type (Mao et al., 2019; Brown et al., 2020), life stage (Zhu et al., 2018), precipitation (Comita et al., 2014) and latitude (Johnson et al., 2012; LaManna et al., 2017). Shade-tolerant species suffered stronger CNDD effect and were on the negative side on the trade-off axis(Fig.4).Shade-intolerant species were scattered on the trade-off axis which showed a random distribution. In closed-canopy forests, shade tolerance species grow slowly with low leaf nitrogen (Brown et al., 2020) and tend to be more conservative to keep stable growth-survival trade-offs in a closed canopy,low light understory environment for long periods(Comita et al.,2009).This trait or strategy may be due to various combinations of the ability not only to survive(tolerate or defend)in low light levels(Krueger et al.,2009), but also to withstand or prevent attack by herbivores and pathogens by allocating resources to storage and defense (Comita and Hubbell, 2009; Queenborough et al., 2013). In contrast, shade-intolerant(light-demanding) species are highly sensitive to shading by neighboring plants,yet can grow quickly in response to high light availability rather than storage or defense(Fine et al.,2006).

In our study, species with different mycorrhizal abundances and in different years generally exhibited different response to CNDD(Qin et al.,2021). Additionally, we accounted for the interaction between shade tolerance and mycorrhizal type,knowing that the EM-associated species in our study plots are typically more shade-intolerant than the AM-associated species (Table S1). In the present study, there was little difference in the number of tree species between the two types(Table S2), which may have resulted in a different result from Bennett et al. (2017). The high abundance of underground roots may lead to a lack of differential accumulation of pathogens around the trees,which in turn may lead to different reactions for AM and EM trees. A study of tropical forests supports the theory that,when conspecific individuals are abundant,soil pathogens suppress seedling recruitment,yet the minimal effect is observed when conspecific individuals are rare due to a lack of pathogens(Liu et al.,2015).That is to say,saplings tend to cluster where there is a high conspecific or low density,making it difficult to produce an effective CNDD.Therefore,larger trees produce CNDD,while smaller trees(saplings and juveniles)are more suppressed,as found by the result in our study that CNDD was strongest at earlier life stages. Compared with the previous results, the CNDD effect exhibits time fluctuation(Kuang et al.,2017;Qin et al.,2021)and the variation of CNDD among species is related to seedling abundance(Zhu et al.,2015a)and may shift over time(Magee et al.,2021).

5. Conclusion

Our approach of demonstrating CNDD from plant traits using tree demographic trade-offs(i.e., growth,and survival) emphasizes that tree survival may respond to neighborhood effects in different ways, where species shade tolerance and mycorrhizal association could affect the degree to conspecifics. However, it is likely missing other commonly measured plant leaves,stem,and root traits such as specific leaf area and wood density that could explain these patterns when only focusing on shade tolerance and mycorrhizal type as mediators of differential susceptibility to CNDD.To better understand the maintenance mechanism of species diversity,it is necessary to examine biotic and abiotic factors that influence tree demographic trade-offs and our findings highlight the importance of plant traits because of their relationship to the underlying ecological processes that lead to interspecific differences in competitive ability and tissue defense.

Funding

This research is supported by the Program of National Natural Science Foundation of China (No. 31971650), and the Key Project of National Key Research and Development Plan (No. 2017YFC0504104), and Beijing Forestry University Outstanding Young Talent Cultivation Project(No. 2019JQ03001).

Availability of data and materials

The tree census data are available from the corresponding author on reasonable request.

Authors' contributions

J.Q.,C.F.,C.Z.and L.G.analyzed the data;J.Q.wrote the manuscript;and Y.G.,L.G.and C.Z.provided comments and other technical support;X.Z. conceived and designed the experiments. All authors contributed critically to the drafts and gave final approval for publication.

Ethics approval and consent to participate

All the authors have approved the manuscript and agreed with submission to your esteemed journal.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We are grateful to the people who helped to collect field survey data.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://do i.org/10.1016/j.fecs.2022.100044.