The role of roots in the resource economics spectrum

The fact that plant traits are correlated in functional trait syndromes within and across biomes has gained considerable importance in ecology (Wright et al., 2004). For example, leaf lifespan is negatively related to specific leaf area (SLA), and further linked via nitrogen (N) content per leaf area to photosynthetic characteristics such as Rubisco, maximum photosynthesis and respiration (Reich et al., 1992). Plants can thus achieve fast growth by rapid assimilation, driven by high leaf N content and a large leaf area with low biomass investment. These large, thin leaves are less mechanically robust and more prone to herbivory than thick leaves and are, therefore, short-lived (Reich et al., 1992). This functional leaf-trait syndrome helps to explain growth–survival tradeoffs, light niche separation and co-existence in tropical forests (Poorter & Bongers, 2006). In temperate forests, the variation in leaf lifespan is more limited, but still explains specialization for shade (Janse-ten Klooster et al., 2007). Belowground plant species traits, however, have not yet been successfully incorporated in the functional trait syndrome conceptual framework, despite several studies addressing this issue (Tjoelker et al., 2005; Withington et al., 2006; Freschet et al., 2010). As biomass allocated to roots varies from 20% in tropical rainforests to 70% in temperate grasslands (Poorter et al., 2012), root traits should be fully integrated into the resource economics spectrum. More specifically, as root production and root lifespan are major drivers in nutrient cycling and water use, our understanding of how these ecosystem processes are shaped will be enhanced by understanding the role of species and their root traits. The study of McCormack et al. (pp. 823–831) in this issue of New Phytologist, is an important contribution to this area. ‘Clearly, the resource economics syndromes that have been widely observed in leaves cannot directly be extrapolated to roots, or may at least not be directly comparable between roots and shoots.’ McCormack and coworkers investigated root lifespan of 12 temperate tree species, among which three Acer species, two Quercus species, Juglans nigra, Populus tremuloides and two Pinus species – growing in monoculture stands in a common garden in Pennsylvania, USA over 4 yr. Minirhizotron tubes (clear acrylic tubes, 3 cm diameter and 45 cm long, installed in the soil at an angle of 30° from vertical, with a camera connected to image analysis software) were used to regularly scan root birth and root death during the growing season. Median root lifespans of individual roots of the 12 tree species were compared with a range of fine root traits determined from destructive harvests, including specific root length (SRL, m root length g−1 dry root mass), root diameter, N, carbon (C) and calcium contents in root tissues, and root respiration. Root lifespans were also correlated to general tree growth parameters such as diameter at breast height (DBH) at a stand age of 10 yr and wood tissue density. Variation in root lifespan ranged from 95 to 336 d among species and a large fraction of the variation was explained by three traits only: DBH, root diameter and N : C ratio (Fig. 1). This finding suggests that relatively simple parameters can be used to estimate an important ecosystem process like root turnover. In contrast to this simplicity, however, are other studies of Withington et al. (2006, on 11 mature tree species in Poland) and Tjoelker et al. (2005, on seedlings of 37 herbaceous and two tree species) that show that a putative functional root trait syndrome may be more complex. These two studies did observe a general correlation between root N (Tjoelker et al., 2005) or N:C content (Withington et al., 2006) and root lifespan, but not with several other root traits, including root diameter and lifespan. Diagram showing the root, stem and leaf traits to be measured in order to develop a whole plant resource economy perspective. The significant traits measured by McCormack et al. (pp. 823–831) in this issue of New Phytologist are presented in red. SLA, specific leaf area; Amax, maximum assimilation rate; Rd, respiration rate; DBH, diameter at breast height; SRL, specific root length. The lack of consistency among studies may be explained by the use of different methodologies in the different studies, as some investigated seedling growth stage (Reich et al., 1998; Tjoelker et al., 2005) and others mature trees (Withington et al., 2006; McCormack et al.). Also, root lifespans have been based on investigations with ingrowth cores (Tjoelker et al., 2005), or minirhizotrons, either calculating root lifespan from the fate of individual fragments (Withington et al., 2006; McCormack et al.) or using the root fragment length as the independent variable (Ferguson & Nowak, 2011). Another explanation for the differences may be that abiotic soil environmental conditions, such as nutrients and water have not been explicitly investigated but do significantly affect root traits and root turnover, and therefore should be explored in future work. Another level of complexity in evaluating the root trait syndrome is the linkage between root traits and the leaf economics spectrum. McCormack et al. did not aim to directly link root traits to the leaf economics spectrum, but a few other studies have explicitly examined this, with mixed results. Tjoelker et al. (2005) found a correlation between leaf and root lifespan, but Withington et al. (2006) did not. Relationships between SLA and specific root length (SRL) were also variable as correlations between SLA and SRL were significant in Wright & Westoby (1999, seedlings of 33 woody shrub and tree species), Withington et al. (2006) and Freschet et al. (2010, 16 woody species and 24 herbaceous species), but not in Tjoelker et al. (2005). Ryser (1996) found a relationship between leaf and root tissue density in five grass species, two traits that were also positively correlated with tissue lifespan. Clearly, the resource economics syndromes that have been widely observed in leaves cannot directly be extrapolated to roots, or may at least not be directly comparable between roots and shoots. To be able to answer the question to what extent plants synchronize their resource capture strategies aboveground and belowground, the functionality of the organs and their main processes need investigation – not only with regard to resource acquisition but also with regard to retention and loss rates, that is, lifespans of roots and leaves. To move beyond the analogy of coupled resource acquisition traits of leaves and roots, the different main functions of the organs need to be integrated in a process-based framework: light acquisition by leaves, acquisition of water and nutrients and the provision of anchorage by roots, and transport of assimilates, nutrients and water by stems and coarse roots (Fig. 1). For example, leaf traits that allow rapid photosynthesis, such as high leaf N and high maximum photosynthetic rate, increase water and nutrient demand. These traits may be functionally coupled with root traits that facilitate these photosynthetic requirements with regard to resource acquisition, such as a large and dense root system with thin roots (high SRL, small diameter) ensuring a large root absorptive surface. High root N content may be associated with a high metabolic activity, but may also be correlated with enhanced root death and decomposition. The biotic component of the soil environment – pathogenic and mutualistic – may be fundamentally different than in air, posing fundamentally different resource economic syndromes in roots compared with leaves. It thus seems that the functionality of different plant parts in the environmental context in particular should be taken into account when examining traits and their interrelationships among tree species. Incorporating wood traits could provide additional insights as stems and branches link roots and leaves, and may pose additional limits to growth of trees and thus forest productivity. The current parallel development of integrative, mechanistic and whole plant models (e.g. Sterck et al., 2011) provides promise for understanding the role of individual traits and traits syndromes for the growth of trees from a whole-plant economic perspective. Improving the understanding of root lifespan and root traits will also benefit our understanding of productivity in mixed forest stands compared with monoculture stands. It has been known for some time that a positive relationship exists between productivity and plant species richness, leading to so called overyielding in mixed stands compared with monocultures. This phenomenon was first investigated in grasslands (Cardinale et al., 2007), but is currently also explored in forests (Brassard et al., 2011; Lei et al., 2012). One of the main explanations for overyielding has been spatial niche differentiation belowground: for example, certain species will root in deeper soil layers in mixtures as compared with monocultures, thus exploring larger soil territories. Belowground overyielding is generally observed in grasslands (Cardinale et al., 2007). However, evidence for niche differentiation as a driving force in grasslands is limited and an experiment with four grassland species even detected greater shallow rather than deeper rooting of the community in mixtures compared with monocultures (Mommer et al., 2010). Other studies have shown that N uptake patterns in grassland mixtures showed less variation than expected based on niche differentiation theory (Von Felten et al., 2012). Experimental studies on forest sites also often, but not always, report belowground overyielding (Brassard et al., 2011; Lei et al., 2012), but evidence for spatial niche differentiation belowground seems more prominent in these studies (Schmid & Kazda, 2002; Brassard et al., 2011). Future studies will need to reveal the exact differences between grasslands and forests in this context, but the larger overall rooting depths of trees compared with grasslands may be key to this difference. If there is the potential for trees to explore over a larger soil depth than grasses, there may be more opportunity for spatial niche differentiation within a mixed forest stand than in a grassland. In summary, the study of McCormack et al. demonstrated significant root trait correlations leading to the possibility of a root trait syndrome similar to what has been observed for leaf traits; however, it is too early to conclude on its generality. We hope that McCormack et al.’s study will stimulate more research effort in this important, yet underexplored research area of functional root trait economics. Information on root lifespan is important to link roots to the leaf economics spectrum of forests, but also to better understand how species richness drives ecosystem functioning in terms of productivity, nutrient cycling and C storage.