More species, more trees: The role of tree packing in promoting forest productivity

Abstract

1 INTRODUCTION

Despite having supplied humanity for millennia with many important goods and services (Brockerhoff et al., 2017; FAO and UNEP, 2020), forests have only recently received large international attention regarding their role in mitigating both climate change and the biodiversity crisis (FAO and UNEP, 2020; Griscom et al., 2017; Pachauri & Meyer, 2014). Many studies have shown that tree species diversity can foster forest productivity and carbon sequestration, resulting in positive diversity–productivity relationships (DPRs) (Brockerhoff et al., 2017; Hooper et al., 2012; Liang et al., 2016). This result is now well-established in the literature and has been corroborated by many methodological approaches relating biodiversity and ecosystem functioning (BEF), including studies relying on forest inventories (Aussenac et al., 2021; Liang et al., 2016; Paquette & Messier, 2011; Ratcliffe et al., 2016; Toigo et al., 2015) or empirical observations (Jucker et al., 2014; Pretzsch et al., 2015), experiments (Sapijanskas et al., 2014; Toïgo et al., 2022; Williams et al., 2017), and simulations with process-based models (Bohn & Huth, 2017; Maréchaux & Chave, 2017; Morin et al., 2011).

DPRs have been assumed to result mostly from species complementarity in resources uptake and use-efficiency (Barry et al., 2019), thus primarily depending on niche partitioning between species. In the case of forests, niche partitioning can occur through root spatial stratification (Cabal et al., 2024), but most evidence concerns light uptake as forest dynamics are generally strongly driven by light availability (Pacala et al., 1996; Rüger et al., 2020), leading to a size-asymmetric competition (Cordonnier et al., 2019; Schwinning & Weiner, 1998). Niche partitioning may lead to a more efficient use of canopy volume in multispecific forests than in monospecific ones and to an increased light interception at the ecosystem level (Guillemot et al., 2020; Rissanen et al., 2019; Williams et al., 2021). This ‘canopy packing’ effect has thus been proposed as a key mechanism explaining the positive effect of species diversity on forest productivity (Morin et al., 2011), and has been evidenced in both temperate (Jucker et al., 2015; Pretzsch, 2014; Williams et al., 2017) and tropical forests (Sapijanskas et al., 2014).

The optimization of canopy packing in multispecific stands is usually explained by two complementary processes: neighbourhood-driven plasticity in crown shape and volume (Guillemot et al., 2020; Jucker et al., 2015; Pretzsch, 2014), and a stronger vertical stratification of tree crowns in the canopy (Morin et al., 2011). Although crown plasticity has received more attention (Guillemot et al., 2020; Jucker et al., 2015; Pretzsch, 2014; Williams et al., 2017), evidence has been provided for both processes, either in observational data (Jucker et al., 2015) or tree diversity experiments (Williams et al., 2021).

Yet, crown plasticity and vertical stratification mostly depend on direct interactions between individual trees. More generally, the hypotheses used to explain BEF patterns in forests are usually based on tree–tree interactions (Trogisch et al., 2021), thus at local scale. This focus may have overshadowed the possible role of complementarity processes on patterns occurring at a larger scale, that is, the scale of the tree community level. This scale, defined as the collection of individual trees within a given area, typically between 10³ and 10⁴ m², is often referred to as the stand in the forestry literature. Here, we hypothesize that a more efficient use of resources due to niche partitioning between species may increase the carrying capacity at the stand scale. In other words, species diversity may raise the maximum stand density, with more trees coexisting in multispecific forests than in monospecific ones. In this hypothesis, we therefore assume that diversity effects on tree–tree interactions at the local scale have also consequences at the stand scale by influencing the number of trees in the community. So far, such a pattern has only been indirectly suggested, for specific mixtures (Pretzsch & Biber, 2016). In addition, we make a step further by hypothesizing that the larger number of trees, resulting from the increased species diversity, may in turn increase stand productivity. This effect, that we call the ‘tree packing effect’ (TPE, Figure 1), is thus a consequence of species complementarity for resource use on spatial coexistence and ecosystem productivity.

FIGURE 1

Open in figure viewerPowerPoint

Theoretical scheme representing the effect of canopy packing (blue path) and tree packing (green path) effects on stand canopy volume and productivity in response to increasing tree species diversity. Processes involved in the canopy packing effect are rather related to tree–tree interactions and do not affect the number of trees per area, but more the individual allometry and functioning of neighbouring trees. Processes involved in the tree packing effect act at the community level, that is, the total number of trees changes at the stand level.

There are several indirect clues in favour of this TPE hypothesis. First, stand density has been known to affect forest productivity for a long time (Forrester, 2014; Reineke, 1933). Second, stand density is usually controlled for in tree diversity experiments (Schnabel et al., 2019; Toïgo et al., 2022; Williams et al., 2017) and in semi-experimental field samplings (Jucker et al., 2015; Pretzsch et al., 2015) that aim at disentangling the effect of species richness on ecosystem functioning through tree–tree interactions. In the same vein, in observational studies, stand density (or proxys for it) has often been considered as a covariate to be controlled to isolate putative effects of tree diversity on productivity, rather than a response variable driving DPRs (Chisholm et al., 2013; Paquette & Messier, 2011; Ratcliffe et al., 2016; Vila et al., 2013). Therefore, the importance of diversity effects on stand density in driving positive DPRs remains largely unexplored (Chisholm & Dutta Gupta, 2023). Yet, if the TPE is confirmed, it implies that a key effect of diversity on forest productivity has been overlooked in BEF-studies.

The TPE thus relies on two components: (i) on average, species richness increases maximum stand density and (ii) this higher stand density enabled by increased species richness promotes forest productivity. To the best of our knowledge, these two components have never been clearly connected and thus tested. Regarding the first one, the positive effect of species richness on stand density has been suggested or indirectly mentioned in several studies (Pretzsch & Biber, 2016; Tatsumi & Loreau, 2023), but has not yet been generally quantified, especially for a large range of tree species and environmental conditions.

The state-of-the-art for the second component of the TPE is also very incomplete. To the best of our knowledge, although former studies have provided some insights about the role of stand density (or proxys for it) on forest functioning in mixed forests (e.g. Brunner & Forrester, 2020; Paquette & Messier, 2011; Ratcliffe et al., 2016), the links between species diversity, stand density, and forest productivity have not yet been clearly and generally depicted. Furthermore, testing for this second component is not straightforward, as higher stand density may be associated with smaller average tree size and/or younger age, possibly leading to a decrease in biomass production per tree. Moreover, understanding the links between tree species diversity, stand density and forest productivity is challenging because they are impacted by many factors, such as climate, soils, stand age, or past management.

Here, we test for the existence of the TPE across a wide range of forest ecosystems and environmental conditions in Europe. Considering the two components of the TPE, we tested (i) whether diverse forests have a larger maximum stand density than monospecific ones, and (ii) whether this can result in a positive effect of species richness on forest productivity. We tested these two components using two separate but complementary analyses, relying respectively on an observational dataset and a process-based simulation experiment.

To test for the first component of the TPE, we analysed the effect of species richness on the maximum stand density in forest plots (N_max [number of trees.ha⁻¹]), defined as the maximum number of trees a plot can sustain at a given developmental stage, which is a well-known rule in forest ecosystems, also called the self-thinning boundary (Forrester et al., 2021; Reineke, 1933). We did so by analysing national forest inventories data from six European countries (Ratcliffe et al., 2016), thus sampling a large diversity of tree species assemblages and environmental conditions.

To test for the second component of the TPE, we used a simulation experiment to explore whether TPE can be involved in shaping DPRs in European forests. Former studies that quantified DPRs in large observational datasets (Liang et al., 2016; Paquette & Messier, 2011; Ratcliffe et al., 2016) did not focus on the link between species richness and stand density and its implications for forest productivity. In fact, evaluating the interactive effects of species richness and stand density on productivity in observational data cannot be done properly because disentangling these effects requires comparing forests with similar tree species composition, in the same environmental conditions but with contrasting stand densities, which is impossible in practice. This is especially the case when focusing on a wide range of species and community composition. Therefore, we tested the significance and strength of the second component of TPE using a validated forest dynamics model in which stand density can be controlled. Relying on functional and demographic processes, such models consider biotic interactions (especially competition for light) and abiotic drivers such as climate (Bohn & Huth, 2017; Morin et al., 2011), and can provide robust predictions of ecosystem composition, structure, and functioning (Maréchaux et al., 2021). Their simulations have been already used to help disentangle the mechanisms behind DPRs especially how climate conditions may modulate these relationships (Bohn & Huth, 2017; Maréchaux & Chave, 2017; Morin et al., 2011, 2018). We thus used the individual-based forest model ForCEEPS (Morin et al., 2021) to test for the effect of maximum stand density (N_max) on DPRs, by simulating forest stands with various species richness levels across the whole range of environmental conditions in Europe, with or without controlling for stand density.