Our study revealed how an applied nucleation technique influences seedling performance mediated by functional traits at both the species and community (cluster) levels in the early stage of restoration (19 months after planting). Despite higher growth rates were found in rows and higher survivorship in clusters at the individual seedling level, these rates did not differ between designs when including the species functional traits, which were stronger predictors. Interestingly, our findings provide partial support for our first hypothesis. As expected, the slow-growing species, characterized by high SDM, high WD, and low SLA, exhibited higher survival rates. However, contrary to our expectations, neighborhood density did not exert a significant influence on species survivorship, except for species with larger seeds whose survival was improved in row plantations (Fig. 2D). Regarding the species growth rate, functional traits representing fast-growing species had a weak influence on it (only potential height was related to growth). Similarly, our second hypothesis at the cluster level was partially validated. Cluster growth (biomass) increased when it was dominated by highly competitive species (i.e., fast-growing) represented by species with lower seed dry mass. However, unexpectedly, a highly functionally divergent community did not promote cluster survival and growth. All relationships at the cluster level occurred regardless of the addition of fertilizer, except for clusters richer in nutrients dominated by larger-seeded species, which increased their stem biomass storage (Fig. 3C).
Low values of SLA (long leaf lifespan), and high values of H (taller species) and WD (hardwood) mostly related to species survival, regardless of neighborhood density. In general, the lack of difference in species survival between seedlings planted in rows and clusters suggests that biotic interactions did not significantly impact the influence of functional traits on survival, at least in the early restoration phases. Our results align with the findings of Lasky and colleagues (2014), who did not find a crowding effect in the relationship between species survival and single functional traits. They found that leaf dry matter content (LDMC) and wood density traits were the main factors that explained species survival. Furthermore, plant functional traits associated with a slow life history (large seeds, long-lived leaves, or dense wood) presented a stronger influence on survival than on growth or fecundity (Adler et al. 2014). These species have been associated with greater resistance to drought, making them well-suited for environments characterized by limited water availability (Oliveira et al. 2021). In this regard, slow-growing species present a “slow-safe” strategy, as they are less vulnerable to hydraulic failure through embolism due to their lower photosynthetic rates and fewer stomatal openings (Oliveira et al. 2021). This hydraulic safety strategy should explain the high level of survivorship of slow-growing species in the early stages of restoration, as seedlings face high transpiration rates.
The functional trait of potential height exhibited an intriguing pattern of positively influencing both growth and survival. Species with greater potential heights seemed to gain a competitive advantage in accessing light and, consequently, carbon (Westoby et al. 2002), driving their enhanced growth rates. In addition, height has been associated with a more extensive root system, facilitating access to water during dry periods (Violle et al. 2009), which may further contribute to the survivorship of taller seedlings. Although we found a weak link between SLA and WD and species-relative growth rates, strong relationships have been found in other studies, in a positive and negative direction, respectively (Kunstler et al. 2016; Poorter et al. 2008; Wright et al. 2010). The higher intraspecific plasticity found in seedling leaves (Havrilla et al. 2021) could have influenced the seedling growth rate and clarifies the lack of an SLA effect. In this regard, the SDM trait could be a better indicator of plant strategies to use than SLA, as it has less variation across the ontogenetic stages of plant life history (Adler et al. 2014).
Our study revealed that only SDM exhibited distinct responses in the two planting designs, indicating that this trait could be linked to the presence of biotic interactions within the clusters (Fig. 2D). As expected, larger-seed species demonstrated an improvement in their survival when planted without neighboring seedlings in rows. A larger reserve of seed resources for seedlings after their emergence may explain their higher survival (Moles and Westoby 2004). However, the presence of immediate clustered neighborhoods decreased these species survival. This result suggests that competition, not facilitation (as expected), led to the assemblage of cluster communities. In such cases, larger-seed species could not coexist in the same niche, explained by the principle of limiting functional (MacArthur and Levins 1967). This result at species level may explain the increased seedling performance at cluster level when they were dominated by fast-growing species (low mean values of SDM). This result is likely explained by their high competitiveness and rapid growth due to their high photosynthetic and transpiration rates, typically from species with more acquisitive strategies (Garnier et al. 2016).
A high functional similarity (low FDiv values) in clusters improved seedling performance, for both survival and growth averages. Low FDiv values indicate that the most abundant species exhibit marked similarity and heightened competitiveness, while high FDiv (low functional similarity) values reflect niche differentiation between species in a community (Mouchet et al. 2010). The biotic mechanisms of facilitation or niche differentiation should improve total resource use in a community and increase productivity (Loreau and Hector 2001) but did not explain seedling performance in such dense cluster communities. Instead, a competitive hierarchy in clusters with a greater abundance of competitive species may exclude less favorable species with functional traits not adapted to this highly competitive community (Chesson 2000), for example, the slow-growing species. Regarding nutrients addition, chemical fertilization improved the cluster biomass stock, but only when they were dominated by slow-growing species (i.e., high SDM values). This result indicates that fertilization may alleviate competition and promote coexistence between those species when soil resources are not limited. Therefore, we rejected our second hypothesis, as we expected that the growth of fast-growing species may be favored in nutrient-rich soil because they are highly competitive (Oliveira et al. 2021).