Plant root and mycorrhizal fungal traits for understanding soil aggregation

Soil aggregation is a key ecosystem process resulting in the formation and stabilization of soil structure, consisting of soil aggregates and the resulting matrix of pore spaces. As such, it significantly alters the environment of plant roots and microbes in a multitude of ways; thus, soil structure provides the basic setting in which mycorrhizas operate and have evolved. Not surprisingly, soil aggregation is important for root growth and for a wide range of soil features and ecosystem process rates, such as carbon storage and resistance to erosion (e.g. Jastrow et al., 1998; Six et al., 2006). The aggregation of soil is a complex process, regulated by a range of abiotic factors (e.g. texture) and mediated by plants and multiple biota groups and their interactions; in spite of this complexity, plant roots and their mycorrhizal symbionts are consistently found to be a crucial force in driving soil aggregation (Six et al., 2004; Rillig & Mummey, 2006; Leifheit et al., 2014). Plant and mycorrhizal fungal species, respectively, differ in their contribution to soil aggregation (Reid & Goss, 1981; Angers & Caron, 1998; Eviner & Chapin, 2002; Rillig et al., 2002; Piotrowski et al., 2004; Six et al., 2004; Duchicela et al., 2012; Graf & Frei, 2013; Pérès et al., 2013), but from previous studies it has not yet become clear what specific root (and mycorrrhizal) traits contribute to this effect. This is likely because studies typically focus on only a limited suite of physiological and architectural characteristics, and the general research focus has traditionally been more on management practices and physico-chemical factors. Therefore, we lack a mechanistic understanding of the role of organisms in soil aggregation. To this end, we argue for a shift from a classical species-based comparative approach, or the mere consideration of summary variables (e.g. root length) to a systematic trait-based approach. Summarizing ecological characteristics of species by means of traits has become an essential tool in plant ecology (Westoby & Wright, 2006), and is increasingly proposed for root–fungal associations (e.g. van der Heijden & Scheublin, 2007; Chagnon et al., 2013; Aguilar-Trigueros et al., 2014). We believe trait-approaches are a promising tool to achieve progress in our mechanistic understanding of soil aggregation from an organismic perspective, to enhance our predictive ability regarding effects of plant and fungal diversity, and to ultimately provide informed recommendations for, for example, agricultural management and environmental restoration. With such data we can then ultimately address questions such as: how can plant and mycorrhizal fungal community data enhance prediction of soil aggregation beyond the consideration of state variables such as soil type, climate and management? Is soil aggregation an intermediate driver of plant and mycorrhizal fungal community processes and thus also an integral part of plant–soil feedbacks? And, on the applied side, can we enhance ecosystem restoration by designing tailor-made mycorrhizal fungal inocula and seed mixes specifically optimized for complementarity in the ability to enhance soil aggregation? In addition to general advantages, we see specific benefits of a trait-based approach applied to the understanding of mycorrhizal and plant contributions to soil aggregation: (1) it will allow testing for the extent of root and mycorrhizal fungal functional complementarity with regard to soil aggregation, and thus whether soil aggregation is another example of an ecosystem process that positively relates to biodiversity like other more well-studied processes such as primary productivity. This is very likely true because soil aggregation is the result of multiple interlocking component processes such as aggregate formation and stabilization (both of microaggregates and macroaggregates) which might be driven by different sets of traits represented in different species. Additionally, soil aggregation may be stimulated through diversity-fueled overyielding of fungi and roots through niche complementarity (also see Table 1); (2) if we can convincingly relate soil structure to traits of plant and fungal communities, this will allow predicting which ecosystems are under most risk of deterioration and also guide towards measures to counter this process. Measure: (loge (root or hyphal lengtht0) − loge (root or hyphal lengtht1)) × (t0 − t1)−1 Rationale: these three traits represent the growth ability of hyphae or roots to encounter building materials of aggregates in soil (primary particles, organic matter particles) and then form macroaggregatesa Proxy: (number of branches) × (root or hyphal length)−1; root or hyphal branching number and angle; alternatively: directly measure ability to engulf an object Rationale: these three traits measure the capability of hyphae or roots to stabilize aggregates depending on their tearing-resistance (i.e. ability to hold material together without breaking apart) and enmeshment potentialb Proxy: average distance between hyphae and associated root surface Rationale: larger distance between root and hyphae allows for increased probability of contact of aggregate building materials with either hyphae or rootsc Proxy: (entry points of hyphae or hyphal density close to root) × (root surface area)−1 Rationale: greater proximity of hyphae and roots increases probability of synergistic/combined effects of roots and hyphae (e.g. cementing agents, entangling) contributing to the stability of engulfed aggregatesc Measure: (surface charge of roots or hyphae) × (root or hyphal length)−1 Rationale: quantity and quality of secreted exudates determine the degree of particles adhering to roots or hyphae and agglutination of particles and aggregatesd Proxy: root water uptake or hyphal water transport (out of root-exclusion compartments) Rationale: these three traits evaluate directly and indirectly the fungal or root impact on soil rewetting capability affecting soil stabilitye Proxy: root or hyphal density in relation to soil porosity Rationale: enmeshment of soil by roots or hyphae lead to soil compression supporting initial aggregate formationf Proxy: stability of artificial aggregates (= soil particles + extracted root or hyphal exudates) Rationale: exudates stabilize aggregates by filling up intra-aggregate pore necks and cracksg Measure: (observation of particles moved by roots or hyphae) × (time unit)−1 Rationale: ability of roots or hyphae to bring soil particles together by moving them, leading to potential aggregation Measure: (root or hyphal growth (length and branching)) × (time unit after disturbance)−1 Rationale: these three traits represent aspects of root or hyphal longevity, resistance and resilience against biotic and abiotic disturbance affecting the potential aggregate stability mediated by root or hyphaeh An approach such as advocated here must begin with a selection of candidate traits related to soil aggregation. Of some traits we already know that they relate to soil aggregation propensity: using correlational approaches (path analysis) to exploit existing environmental gradients in landscapes, coarse-level indicators such as root biomass (or often root length) and mycorrhizal fungal hyphal length in soil have been identified as significant determinants (e.g. Miller & Jastrow, 1990; Rillig & Mummey, 2006). Thus, clearly productivity-related characteristics like root and hyphal density are traits to be considered. However, when such measures of root and hyphal abundance were compared to soil aggregation at a finer scale, for example in small field plots (e.g. Rillig et al., 2002) or pot experiments (e.g. Piotrowski et al., 2004), there was generally a low match, prompting the consideration of more specific physiological or architectural traits. In order to select these specific traits, we have first divided soil aggregation into the component processes soil aggregate formation and the stabilization of existing aggregates, and then considered measurable plant and fungal traits that are likely to be variable, and may mechanistically relate to either or both of these processes. For a full list see Table 1, and later we will further explain why these traits were selected and which difficulties may be encountered. Regarding aggregate formation and stabilization, even though these processes will occur simultaneously in ecosystems, these may be executed by different organisms expressing different traits. We define soil aggregate formation as the initial binding together of particles, whereas soil stabilization is the process that renders the aggregate increasingly resistant to the application of disintegrating forces, such as water penetrating into pores. If an aggregate is nonstable it will disintegrate; it has become clear that formation, stabilization and disintegration occur in a dynamic fashion in soils (e.g. Six et al., 2004). For the formation of soil aggregates, we consider traits to be important that are related to the likelihood to encounter soil particles (e.g. fineness of roots and hyphae, their extent and spatial distribution) and to bring soil material together (ability to move, entangle and engulf). In addition, exuded materials from roots and hyphae should play an important role in the initial formation of aggregates (Six et al., 2004). For stabilization of existing aggregates, traits related to persistence are hypothesized to be more important, such as tensile strength or longevity of roots and hyphae. The ability to render surfaces hydrophobic could also be important, as well as the release of cementing agents (Oades & Waters, 1991). We emphasize that these are hypotheses at this stage; only controlled experiments employing a range of plant and fungal species will reveal which of these traits are in fact of explanatory value at all, and for which component process. We realize that some of these traits will be relatively easy to measure, and have also been proposed as general functional traits (see Van der Heijden & Scheublin, 2007); others will be quite a challenge, and for some we propose proxy traits that will be easier to capture (Table 1). Clearly, much innovation is still possible once this research effort is under way. Another challenging aspect of linking single trait values to strains/species is trait variability (e.g. Cordlandwehr et al., 2013), which can be caused either by genetic variability at the species level or by phenotypic plasticity within the same genotype. Intraspecific diversity can have a profound influence on trait expression in mycorrhizal fungi (Johnson et al., 2012; Angelard et al., 2014) as well as in plants (e.g. Kichenin et al., 2013). Nevertheless, it remains yet to be seen if root and mycorrhizal fungal traits are more variable than other plant traits. Apart from this challenge, investigating this very plasticity could yield further insights; for example, some root traits have been proposed as highly relevant response traits in ecological studies due to their high plasticity (Ryser, 2006; Valverde-Barrantes et al., 2013), and if these include traits responsible for soil aggregation we might better understand under which circumstances plants are more likely to contribute to soil aggregation. For much of what we propose here, dedicated experiments are necessary, for example, for the separation of formation and stabilization of aggregates, or for comparing different root systems or fungal isolates under otherwise identical conditions. Nevertheless, several other approaches are possible to make inroads. For example, some trait data are already available because they have been collected for other purposes. Nevertheless, at present, plant trait databases (e.g. TRY, Kattge et al., 2011) are relatively poor in terms of coverage of root traits, in particular root traits that would be important in our specific context (Table 1). So far, comparable mycorrhizal fungal trait databases do not yet exist. Thus one important research focus should be the collection of root and mycorrhizal fungal traits for a number of species and linking these with experimentally measured soil aggregation, ideally dissected into component processes. These could be complemented by observational approaches, consisting of measuring community-average root traits in the field, which are linked to soil aggregation (here it would not be possible to distinguish formation and stabilization). Such data could also be particularly valuable for verifying predictions derived from knowledge of species-level data. Using traits can increase our fundamental understanding of the intricate relationship between plants, symbiotic fungi, and their immediate environment which is the soil aggregate. However, knowledge of root and mycorrhizal fungal traits could also have great applied significance. Such data could give rise to innovations such as tailored seed mixes (or fungal inoculum mixes) for grassland restoration which maximize trait coverage in terms of soil aggregation within the available plant species pool, as well as setting priorities for conservation efforts through predicting which ecosystems are most prone to degradation in light of invasive species and imminent global change. Likewise, better information on plant traits could be used to foster crop breeding for sustainable agriculture, or for agroecosystem management to enhance soil stability (e.g. by selecting cover crops also for complementary trait values).