Integrating microbial ecology principles into growing media formulation for reliable biological functions

1 Introduction

Horticultural growing media (GM) are increasingly expected to support not only plant growth, but also biological functions such as organic matter turnover, nutrient mineralization and disease suppression (Raviv et al., 2019). Yet these functions remain variable and difficult to predict across formulations, crops and production systems, as GM are designed and evaluated mainly through physical and chemical properties, while biological functioning is treated as secondary or assumed to transfer between systems.

As horticulture moves toward peat-reduced formulations, increased use of organic side-stream materials, reuse of spent media and more automated production systems (Blok et al., 2021; Järvenpää et al., 2021) variability in resource quality, microbial legacies and environmental constraints will increase, making empirical formulation increasingly unreliable. Predictive design therefore requires integrating microbial ecology from the outset.

Accordingly, we argue to treat GM as engineered, partially self-organizing microbial ecosystems in which substrate chemistry, stoichiometric balance and environmental filtering shape microbial assembly and function where biological performance, i.e., the context-dependent expression of microbial functions and emergent microbiome-mediated processes in the GM, thus arises from formulation and management rather than inoculation alone. We advocate a shift from empirical blending to ecology-based GM design. This perspective paper develops a conceptual framework with practical implications. Key terms are defined in Supplementary Table 1.

2 An ecology-based framework for biological performance in growing media

For ecology-based GM design, the key issue is not microbial presence per se, but how GM properties and environmental conditions shape microbial assembly and function. This is especially important because horticultural GM span a broad gradient from mineral media (e.g., rockwool, pumice, perlite) to organic media based on plant fibers or biologically transformed side streams. So, biological properties should not be assumed to transfer across GM classes (Raviv et al., 2019).

2.1 Biological functionality in growing media is governed by a small set of interacting controls

Reliable biological performance in GM depends on a limited number of ecological drivers that determine microbial outcomes across systems, namely:

• resource quality, particularly the quantity and chemical composition of organic matter (Asada et al., 2022; Čapek et al., 2021),
• stoichiometric balance, especially the relation between carbon availability and nutrient demand (Li et al., 2021),
• environmental filtering, including pH, water availability, aeration, salinity, temperature and physical structure (Fierer and Jackson, 2006; Liu et al., 2023; Rousk et al., 2010; Silva-Sánchez et al., 2019).

These factors jointly regulate microbial assembly, metabolic allocation and carbon partitioning between growth, maintenance, respiration and extracellular enzyme production. These fundamental microbial ecology concepts are well-established in soil microbiology. Concept-based mapping has been proposed for horticultural GM (Alsanius et al., 2024), but microbial ecological concepts are rarely central to formulation or study scope (see Supplementary Table 3). Processes explainable through ecology are described in some contexts, e.g., microbial successions during composting (e.g., Suárez-Estrella et al., 2025), priming (e.g., De Haas et al., 2025) and environmental filtering (e.g., Naik et al., 2026) yet are seldom validated in GM. The classical paper by Postma et al. (2000), which explained disease suppressiveness in a rockwool–cucumber–Pythium aphanidermatum pathosystem using a Koch’s postulate-like approach, demonstrates that ecological concepts can be translated into GM research as testable, system-specific hypotheses.

Ecological interactions are especially important in GM because these systems differ strongly in biological starting conditions and substrate composition in the sense of primary organic carbon pool, functioning as an energy source and nutrient reservoir sustaining microbial growth, maintenance and enzyme production. In mineral or low-organic GM, microbial activity is commonly constrained by low intrinsic carbon availability and strong dependence on external nutrient and carbon inputs (Alsanius and Wohanka, 2019). In contrast, organic GM contain larger and more heterogeneous resource pools, but the resulting biological activity depends on how carbon chemistry and nutrient availability align with microbial demand. As a result, biological properties cannot be inferred from “organic” or “mineral” status alone and should not be assumed to transfer between GM classes (Grunert et al., 2016).

For GM design, the practical implication is that microbial processes must be interpreted in context. A formulation rich in labile and balanced resources may support rapid microbial growth and nutrient turnover, whereas carbon-rich but nutrient-poor substrates may promote nutrient immobilization, slower turnover and stronger competition for limiting resources. Likewise, pH, oxygen supply and water status determine whether these strategies can be expressed under production conditions. Biological performance in GM is therefore a product of substrate chemistry × stoichiometric context × habitat conditions, not of inoculation or material choice alone.

Opposite to soil, GM are manufactured and managed along a defined use chain. Blending, storage, transport, potting, crop establishment and growth (Green et al., 2006; Khalil and Alsanius, 2001; Rosberg et al., 2014), management (e.g., irrigation, fertigation) (Li et al., 2023, Li et al., 2016) and reuse (Vandecasteele et al., 2020) can each alter the GM’s ecological context. Furthermore, root growth alters temperature and oxygen availability and supplies labile carbon via rhizodeposition, stimulating microbial activity and contributing particulate carbon through root turnover (“priming”) (Kuzyakov et al., 2000). Consequently, roots exert localized control over microbial community composition and function, depending on plant species, developmental stage and GM properties (Sasse et al., 2018). Biological properties in GM should therefore be understood as state-dependent outcomes that change over time, especially in organic GM, where degradation, nutrient dynamics and microbiome trajectories may shift during storage and cultivation.

2.2 Predictive framework for biological performance in growing media

A predictive framework for GM does not require importing the full catalog of microbial ecology into horticulture. We propose that four linked ecological drivers are sufficient to structure biology-oriented GM design, namely (i) resource quality and stoichiometric constraints, (ii) environmental filtering, (iii) microbial strategies and community assembly and (iv) temporal system development (Figure 1). Together they explain why microbial functions vary between GM formulations and why biological performance often remains inconsistent when formulations or production conditions change. In detail,

• resource quality and stoichiometric constraints define the metabolic landscape in which microbiota operate (Hall et al., 2011). Organic matter chemistry and C:N:P balance determine whether microbes allocate resources to rapid growth, nutrient acquisition, maintenance or decomposition of structurally complex substrates (Fierer et al., 2007; Huang et al., 2021; Li et al., 2021; Paillat et al., 2020; Sinsabaugh et al., 2016). Bulk descriptors alone are therefore insufficient to predict nutrient immobilization, mineralization or decomposition rate.
• environmental filtering links formulation to function by determining which microbial traits can persist and be expressed. In GM, pH, aeration, moisture dynamics, salinity and structural organization are not background variables. They actively select microbial phenotypes and thereby regulate metabolism, interaction networks and community stability. Because these filters are modified by manufacturing and crop management, biological performance must be interpreted within clearly defined system boundaries.
• microbial strategies and community assembly explain how microbiota respond to that landscape. Fast-growing opportunists, slower decomposers, stress-tolerant taxa and interacting consortia do not contribute equally to biological performance. Their relative importance depends on resource availability, disturbance regime and competition (Bohacz, 2018; Geyer et al., 2016; Kallenbach et al., 2019; Siles et al., 2024). This also helps explain why inoculants perform inconsistently and why similar physicochemical formulations may diverge biologically.
• temporal system development links to the three previous ones and is particularly essential given GM change through storage, cultivation and reuse (Grunert et al., 2014, Grunert et al., 2016; Khalil and Alsanius, 2001; Vandecasteele et al., 2020, Vandecasteele et al., 2021). Biological functionality must therefore be assessed as a trajectory.

These dimensions can be used to interpret three biological outcomes that are especially relevant for GM engineering. Organic matter turnover and nutrient dynamics (mineralization/immobilization) are biogeochemical processes directly linked to microbial metabolism and carbon allocation (Philippot et al., 2024). Disease suppression, in contrast, is an emergent regulatory property that depends on pathosystem-specific interactions, microbial network structure and resource context (Postma et al., 2005, Postma et al., 2009; Steinberg et al., 2019). As such, disease fails to develop despite favorable conditions for the pathogen and a susceptible host in response to either high microbial density and competition (general suppression) or defined antagonists or metabolites targeting particular pathogens (Weller et al., 2002). It should not be treated as an inherent property of organic GM, but as a context-dependent outcome that must be validated for the relevant pathosystem-GM combination.

Taken together, this framework changes how biological properties in GM should be approached. It shifts attention from descriptive microbiology and post hoc inoculation to design logic: which ecological drivers matter most, how they interact and which measurable indicators are needed to evaluate whether a medium moves toward the intended biological state. This is the basis for more reliable formulation, monitoring and future decision support in controlled horticulture.

2.3 Microbial inoculants must be evaluated as ecological interventions, not standalone inputs

Microbial inoculants are often marketed as direct solutions for plant growth promotion, nutrient mobilization or disease control, but mainly target soil use (O'Callaghan et al., 2022). An important exception is the review by Mourouzidou et al. (2023) that synthesizes existing literature on the role of inoculants in GM and establishes the interaction variability between inoculants and GM.

Inoculants do not enter empty systems. They are introduced into already structured, complex microbial communities and resource environments. Their success therefore depends not only on intrinsic traits, but also on niche availability, resource accessibility, timing of introduction, priority effects, physicochemical compatibility (pH drift, oxygen decline, resource pulses) and interactions with resident microorganisms (e.g., phage/bacterial/fungal/oomycete interactions).

In this perspective, inoculants should be treated as ecological interventions. They may behave as invaders (Mallon et al., 2015a, Mallon et al., 2015b; Shah et al., 2010), but they may also function as mutualists, metabolic complementers or trajectory-shaping taxa within an assembling community. This broader view matters because performance in GM is not determined solely by persistence, but by whether the inoculant modifies community structure or function in ways that support the intended biological outcome (Mallon et al., 2018). An inoculant that fails to dominate may still influence decomposition, nutrient cycling or disease suppression if it alters interactions within the resident microbiome.

Consequently, this has direct implications for GM formulation. Inoculants should not be selected independently but aligned with the GM#s resource profile, its environmental filters and the resident community they will encounter. In practice, GM-specific compatibility testing, persistence assays and timing-sensitive introduction strategies are more informative than broad claims of product efficacy across GMs. The relevant question is not whether an inoculant “works” in general, but under which GM conditions, at which production stage and for which target function it can contribute reliably.

2.4 Requirements of an ecology-driven growing media engineering framework

If biological performance is to become a design variable in GM, ecological reasoning must be translated into formulation and monitoring practice. Here, predictability refers to the ability to anticipate a target biological function within defined system boundaries from a known set of ecological drivers and indicators. Operationally, it is expressed as reproducibility within a GM context and as interpretable performance across contrasting formulations, production stages or crop contexts when assessed using standardized descriptors, process-based measurements and explicit validation steps (Table 1). At minimum, this requires four changes:

Table 1. Matrix linking growing media (GM) state, target biological function, minimum indicator sets and intended design or monitoring use. The table summarizes how biological assessment in GM may be matched to system state across the manufacturing-to-cropping chain. For each GM state, the minimum indicator sets required to assess organic matter turnover, nutrient mineralization/immobilization and disease suppression or inoculant performance differ according to the intended use of the information, such as constituent screening, formulation design, quality control, crop-stage monitoring or reuse decisions. The matrix emphasizes that biological indicators are not universally informative, but must be selected in relation to ecological context, target function and stage of GM use. NB! Displayed indicators are suggestions extrapolated from microbial ecology concepts and must be assessed and validated in GM contexts (see also Supplementary Materials S2–4).

CO₂ formation in incubation assay; ²CO₂ production; ³CO₂ production over time; ⁴likelihood for microbial retention of available nutrients.

• GM should be characterized using descriptors that capture biologically relevant resource quality, not only bulk physical and chemical traits.
• Stoichiometric conditions should be reported in a way that allows interpretation of likely microbial nutrient limitation and turnover dynamics.
• Biological outcomes should be assessed across the GM use chain rather than at a single sampling point (temporal development).
• Inoculant performances and disease suppressive claims should be validated under GM-specific ecological conditions rather than generalized across GM classes.

This also implies clearer minimum indicator sets (Table 1). Mentioned indicators are suggested by analogy from microbial ecology, but have not yet assessed for GM. For decomposition and nutrient dynamics, indicators should capture resource pools, microbial biomass and process rates. For disease suppression and inoculant performance, indicators should capture community structure, community dynamics, interaction context and the relevant pathosystem. Although this perspective article cannot resolve all methodological questions, predictive GM design depends on standardized indicators, explicit system boundaries and biologically informed validation strategies. At present, the oxygen uptake rate, a CEN standard method (CEN, 2020), and sometimes C/N ratio, are determined in organic GM, but, they are not integrated with other descriptors to enable predictive modeling. Microbial community composition is assessed at different levels of sophistication, but on functionality is often missing. The corresponding concepts and candidate indicators are summarized in the Supplementary Text S2, Supplementary Figure 3 and Supplementary Table 4.

A practical ecology-driven GM engineering framework should therefore follow four validation steps:

• Define system boundaries, including GM type, stage in the production chain, crop context and management regime.
• Define target biological function, such as organic matter turnover, nutrient mineralization, disease suppression or inoculant performance.
• Select a minimum indicator set appropriate to that function, combining substrate descriptors, microbial indicators, and process-based measurements.
• Test performance across contrasting GM formulations and time points (temporal development) to determine whether the intended biological outcome is reproducible and transferable.

Such stepwise approach is essential if biological functionality is to move from descriptive assessment to predictive GM design (see also the illustrative operationalization of disease suppression hypotheses in Supplementary Table 5).

3 Discussion

This perspective’s central implication is that biological performance in GM cannot be improved reliably by inoculation or material substitution alone. It depends on how resource quality, stoichiometric balance, physiochemical filtering and microbial community assembly are combined within a defined GM context. Treating GM as engineered, partially self-organizing microbial ecosystems therefore shifts formulation from post hoc description to design and management of ecological conditions that make target functions more predictable.

This shift is especially relevant because GM are becoming more heterogeneous with increased variation in organic matter quality, microbial legacies and environmental constraints. Under these conditions, assumptions of transferability across GM types become increasingly unreliable. Results from one GM class, production stage or crop context should therefore not be generalized without explicit testing across contrasting GM formulations and clearly defined system boundaries.

A narrow path separates overgeneralization from excessive detail. This is particularly crucial for disease suppression. Disease suppression should not be treated as an inherent property of organic GM and inoculant establishment should not be inferred from taxonomic presence alone. Both are context-dependent outcomes shaped by pathosystem identity, substrate chemistry, environmental conditions and community history. More broadly, many ecological principles used here are adapted from soil microbial ecology, but their relevance in GM depends on how well they are redefined within manufactured, managed and temporally dynamic substrates. Clearer system boundaries and more explicit validation steps are therefore essential.

3.1 Priority 1: standardization of biologically relevant indicators

At present, GM are still reported mainly through physical and chemical properties, while indicators most likely to explain microbial functioning remain inconsistently measured. A more predictive framework requires routine reporting of variables that capture organic matter quality, stoichiometric balance, and environmental filtering, including fractionated carbon pools, C:N:P ratios, pH, water status, aeration and related process indicators. Without this, synthesis across studies remains limited and biological performance continues to appear more context dependent than it may actually be.

3.2 Priority 2: cross-GM validation

Claims about decomposition dynamics, nutrient mineralization, disease suppression or inoculant success should be tested across gradients in GM composition rather than within a single formulation type. This is particularly important where mineral and organic GM differ strongly in carbon availability, buffering capacity, degradation dynamics and microbial legacy effects. Cross-GM validation is needed to identify where ecological principles are transferable and to determine where formulation-specific mechanisms dominate and require separate design strategies.

3.3 Priority 3: GM-specific ecological compatibility testing for microbial inoculants

The inconsistent performance of inoculants is not only a product-quality issue; it is also an ecological issue. Inoculant success depends on whether the introduced organism encounters suitable resources, compatible habitat conditions and sufficient niche opportunity within the resident community. Inoculants should therefore be evaluated in relation to the GM targeted for introduction, the stage of the production chain, and the target function they are expected to support. Assays based on persistence, niche overlap and priority effects are likely to be more informative than generalized efficacy claims across GM classes.

3.4 Priority 4: temporal sampling across the manufacturing-to-cropping chain

Biological properties in GM are not fixed but change during the manufacturing and management. One-point measurements thus provide only a partial view of system function. Time-resolved sampling is particularly important for both understanding decomposition, nutrient immobilization–mineralization dynamics and shifts in disease risk and for interpreting inoculant persistence and microbiome succession. A trajectory-based view of GM biology is therefore essential if formulation is to become predictive rather than reactive.

These four priorities jointly urge for an additional priority:

3.5 Priority 5: integration with digital monitoring and decision-support tools

In the future, GM design must be compatible with industrial scalability, automation and closed-loop control in controlled horticulture. Future work should therefore connect microbial indicators to routinely monitored variables, such as moisture, oxygen status, electrical conductivity, pH and temperature. This is important for identifying interpretable indicator sets that can support early warning, adaptive fertigation, inoculant timing and GM management decisions. By integrating digital monitoring and decision-support tools, biological properties can gradually move from an uncontrolled source of variability to a manageable design variable in precision horticulture.

In summary, the next phase of GM research should move beyond asking whether biological functions occur and instead ask under which ecological conditions they can be designed, measured and stabilized. Standardized indicators, cross-GM validation, inoculant compatibility testing, temporal sampling integrated with digital monitoring define a realistic agenda for transition. If implemented, this agenda supports a shift from empirical formulation toward ecology-based GM design and provides a stronger foundation for biological performance in soilless horticulture.

4 Conclusions

• GM are engineered microbial ecosystems, not passive plant supports.
• Biological performance emerges from resources, stoichiometry and environmental filters.
• Soil ecology must be adapted to GM-specific conditions and dynamics.
• Predictive GM design requires clear boundaries, indicators and validation.
• Progress depends on standardization, time-series data, digital tools and transparency in data collection and access.
• Future research should address these issues to establish ecology-based GM design and reliable, predictable biological performance.

Generative AI statement

The author(s) declared that generative AI was used in the creation of the manuscript to condense the text. The text condensation was performed to fit the text within the word limit for Frontiers. It was done without compromising on intention, level of detail nor scientific accuracy. Suggested text condensation was critically revised by the authors.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fhort.2026.1794701/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Xiao Yang, Chinese Academy of Agricultural Sciences, China

Reviewed by: Kumar D. Gahlot, Umeå University, Sweden

Sashidhar Burla, ATGC Biotech Private Limited, India