Rethinking tillage as a strategic practice in agroecological farming systems with carbon, climate, and pest management trade-offs

Abstract

Agroecological transitions face a dual imperative: reducing pesticide reliance and minimising soil disturbance for climate mitigation. However, the blanket promotion of no-till practices may hinder ecological innovation by overlooking real-world trade-offs. This article re-examines tillage as a strategic practice in pesticide-free systems, especially for weed and pest control and cover crop management. We synthesise recent findings on soil organic carbon (SOC) dynamics and greenhouse gas emissions, showing that tillage impacts SOC distribution in the soil profile more than stock, and that nitrous oxide (N₂O) emissions can offset potential SOC gains under no-till. SOC stocks are mostly controlled by the quantity and diversity of carbon inputs. We argue for an integrated, systems-based agroecological approach that assesses tillage within the broader framework of carbon inputs, ecological functions, and greenhouse gas balances. Reframing tillage is essential to design resilient, low-input, climate-positive farming systems.

Introduction

The transition towards agroecological farming systems, based on enhanced mobilisation of natural processes and sustainable food systems, presents unprecedented challenges and requirements for agricultural practices^1,2,3. On one hand, there is an increasing pressure to drastically reduce the use of synthetic pesticides due to concerns over biodiversity loss, environmental health, and human exposure risks^4,5,6. On the other hand, climate mitigation goals are driving the adoption of low-emission farming strategies, with no-till practices being promoted to enhance soil organic carbon (SOC) sequestration⁷. They are currently gaining popularity as practitioners become aware of the important role of agriculture in mitigating climate change. This dual imperative to reduce pesticide dependence and minimise soil disturbance through tillage poses a fundamental dilemma for agronomists and farmers alike to maintain the productivity of agroecological farming systems.

Soil tillage, in particular ploughing, is historically criticised for its association with carbon loss and erosion, but remains a powerful tool for managing soil-borne pests, weeds, diseases and cover crop incorporation in the absence of chemical controls^8,9,10. However, such practices are often viewed with suspicion in climate policy circles. Trade-offs about these different aspects of agroecological transition are poorly documented, and the net combined effect of all soil and crop management practices on climate mitigation outcomes is seldom discussed.

This paper argues that the current binary discourse of “tillage bad, no-till good” fails to capture the complex trade-offs that define real-world farming systems, as it overlooks the wide range of potential tillage intensities, frequencies, and modes of implementation within a cropping system. Building on a synthesis of cutting-edge understanding of SOC persistence complex processes occurring at the soil pore scale and their controls (Fig. 1), and through a critical synthesis of recent field experiments and meta-analyses, we explore the interplay between tillage, SOC sequestration dynamics, greenhouse gas (GHG) emissions, and pest regulation services. Our central claim is that tillage should not be considered as a universally harmful practice, but rather as a situational measure within a continuum of practices differing in intensity and timing (Fig. 2), strategically implemented to manage soil and ecosystems. We advocate for an integrated approach that weighs the total GHG balance alongside agronomic and ecological functions of agroecological systems. Reframing tillage in this way is essential for designing resilient, climate-positive cropping systems that can thrive without synthetic pesticides.

Fig. 1: Soil organic carbon (SOC) storage and greenhouse gas (GHG) emissions are emergent system properties of the processes occurring at the soil pore scale, themselves controlled by soil management strategy.

Adapted from refs. ^99,100,101. At the soil pore scale, organic materials (i.e., plant and animal residues) are continuously biodegraded by microbial decomposer communities through the action of exoenzymes, leading to the formation of small-molecule compounds (e.g., monomers). Complete biodegradation results in the mineralisation of carbon (C) through microbial respiration and the release of inorganic nitrogen compounds, ammonium (NH₄⁺) and nitrate (NO₃⁻). Microorganisms assimilate both small organic and inorganic compounds to support their growth. Soil fauna strongly regulates microbial population and activity through bioturbation. Nitrous oxide (N₂O) is produced through various microbial processes, notably nitrification in oxic microsites and denitrification, a main source of N₂O, in anoxic microsites. The enhanced oxidation of biologically transformed organic materials increases their solubility in water, reactivity, and facilitates their persistence via reversible interaction with mineral surfaces, and physical protection through incorporation into soil aggregates. Thus, microorganisms play a dual role during decomposition: they lose part of the C through respiration, but they also drive the formation and persistence of SOC from the remaining part. C persistence increases initially when substrates are inaccessible to decomposers, although such inaccessibility is often transient. More sustained persistence occurs after C has been transformed into microbial-derived compounds that display a strong affinity for soil minerals, thereby promoting mineral-associated organic matter formation. Soil C persistence in soil is controlled by the high molecular diversity of organic materials that require diverse metabolic investments from microorganisms to utilise these molecules, the spatial inaccessibility of organic substrates due to soil structural heterogeneity, the decreasing probability that decomposers meet substrates or exoenzymes to efficiently degrade them, and the abiotic conditions that influence the process intensities. For example, temperature and water availability enhance exoenzyme activity and mobility, dioxygen and nutrients are required to support the life of decomposers, while the dioxygen depletion in soil microsites promotes a shift to denitrification. Elevated pH promotes nitrification and organo-mineral interactions. The application of agricultural management practices that increase both the quantity and diversity of organic inputs returned to the soil enhances molecular diversity. Practices that promote soil life further stimulate the production of microbial-derived compounds and the formation of biogenic structures. In addition, reducing the frequency and intensity of soil mixing by tillage helps prevent faunal and fungal depletion, soil erosion, and limits the accessibility of organic substrates to decomposers. Altogether, these practices are recommended to positively control soil C persistence. Their implementation and repetition over time constitute the operational expression of the soil management strategy defined by the practitioner. The SOC sequestration and GHG emissions are the emergent ecosystem properties resulting from the interaction between decomposer communities, the diversity of their organic substrates, and the heterogeneous and dynamic soil environment. Transient peaks in CO₂ and N₂O emissions—arising from the activity of biotic processes at the soil pore scale—do not necessarily indicate irreversible impacts on the system. However, the dynamics and trajectories of these emissions should be continuously monitored to ensure that soil and system management strategies remain aligned with biosphere-friendly objectives. In the soil pore-scale diagram, the lowercase letters ‘a’, ‘m’, ‘n’, ‘d’, and ‘b’ denote the following biotic processes, respectively: assimilation, mineralisation, nitrification, denitrification, and bioturbation. Solid arrows represent biotic transfers, while dashed arrows indicate predominantly abiotic transfers. Line thickness corresponds to relative transfer rates, with thicker lines representing faster processes.

Fig. 2: System-level effects of soil tillage and C-friendly practices on cropping system performance (SOC storage, GHG emissions, pesticide use), accounting for interactions and synergies among practices rather than isolated effects.

C-friendly practices increase the quantity and diversity of organic C inputs, which is recognised as the most efficient way to increase SOC storage. They can also reduce GHG emissions by replacing mineral N inputs with biological N fixation and organic amendments. The signs + and – relate to the effect of soil tillage on the considered impact. The direct effect of soil tillage on SOC storage is uncertain, but is generally small or even negligible when considering the entire soil profile. However, frequent tillage can indirectly affect SOC storage by reducing the growth period of crops, particularly cover crops. No-till farming reduces GHG emissions linked to fossil fuel combustion, but these emissions generally account for only a small part of total GHG emissions from cropping systems. Furthermore, reducing soil tillage has contrasting effects on N₂O emissions, depending on the context. Therefore, the overall effect of tillage on total GHG emissions is particularly uncertain. No-till systems generally require more pesticides, particularly herbicides. Soil tillage is a strategic tool to reduce pesticide dependency.

Tillage as a pest management tool in low-input systems

In the pursuit of cropping systems with low pesticide inputs, tillage is regaining attention not as a relic of conventional farming, but as a practical and sometimes indispensable tool for pest regulation. While conservation agriculture and no-till systems are often framed as ecological best practices^7,11, their ability to suppress weeds, pathogens, and invertebrate pests without chemical inputs remains highly variable. Mechanical interventions, including shallow tillage, strategic ploughing, false seedbed preparation, and mechanical weeding, are integral to managing these challenges in low-input systems.

Long-term experiments and on-farm studies have highlighted the critical role of tillage in weed seedbank management under a wide range of agropedoclimatic conditions^12,13,14,15. A substantial literature documents the major challenge of weed management in strategies based on no-tillage in contrasted pedoclimates and types of cropping systems, including corn-based systems in the US^16,17, wheat-based systems in Canada¹⁸ or in Europe¹⁹, or Africa²⁰. No-till increases the reliance on herbicides compared to systems based on conventional tillage, and particularly on glyphosate⁸, and the risk of development of weed biotypes resistant to herbicides²¹. Weed biotypes resistant to glyphosate have been described for many species throughout the globe and could become a major challenge for cropping systems highly reliant on glyphosate²². Cropping systems based on shallow tillage only, without soil inversion, also often face challenges with weed control^16,17. Model simulations of blackgrass (Alopecurus myosuroides), a problematic weed in European cereal systems, demonstrate that even with 95% herbicide efficiency, population escape is likely without additional mechanical control²³. This is particularly relevant in pesticide-free systems, where herbicide options are absent. False seedbeds and shallow cultivation can significantly reduce weed pressure by targeting germination-sensitive stages, often outperforming biological or rotational alternatives in short timescales.

Data from commercial farms engaged in the French ECOPHYTO DEPHY network further underlined the role of tillage in shaping integrated weed management strategies. By using regression tree analysis, Lechenet et al. ²⁴ identified that farms with contrasting pesticide use indexes often differ more in their rotation strategies and tillage intensity than in soil or climatic constraints. In certain livestock-integrated systems, reduced tillage may be feasible without compromising weed control, but in arable contexts, its removal frequently necessitates costly or unsustainable chemical alternatives.

Moreover, recent findings from the INRAE CA-SYS experiment in Dijon, north-east France, show that no-till systems without synthetic pesticides can fail within a few years, especially under high weed or pest pressure. In contrast, systems using occasional tillage to manage weeds or slugs maintained both productivity and ecological goals²⁵.

No-till is also most often associated with increases in the risk of soil-borne diseases, compared to cropping systems with conventional tillage, and many cases were documented in cereals for rhizoctonia root rot^{26,27,28,29,30}, fusarium crown and foot rot^26,29,31, tan spot³¹ and pithium root rot²⁶. For a few diseases, however, impact of no-till could be contrasted over different experimental sites (e.g., take all due to Gaeumannomyces graminis), and in a few cases, soil-borne diseases were reported to have a weaker incidence in no-tilled fields (Gaeumannomyces graminis and Pratylenchus neglectus²⁹), eventually as a consequence of higher abundance and diversity of soil fungal communities in no-till fields, compared to fields tilled conventionally, with more abundance of antagonistic fungal taxa that could play the role of biocontrol agent against root pathogens³².

Occasional tillage (also referred to as ‘strategic tillage’) has been suggested as an option to solve challenges of weed management and soil compaction in no-till systems. Many experiments tested different occasional tillage and assessed the impacts on weed communities (Table 1). All studies (either meta-analyses or individual experiments) found strong benefits of occasional tillage for the regulation of weed infestations in the first year after tillage. The evolution of weed communities in subsequent crops (from year 2 after occasional tillage, i.e. when coming back to no-till and direct seeding) displayed contrasted patterns. Mechanistic modelling of weed demography as an effect of cropping systems demonstrated the cumulative benefits of occasional burying of weed seeds, every four or eight years, to ensure a better regulation of weeds in the long term, compared to continuous no-till, and drastically reduce the risk and speed of development of weed biotypes resistant to herbicides³³.

Table 1 Impacts on weeds of one or several tillage operations following several years of continuous no-till: a synthesis of field-based studies

Soil tillage is therefore a key component of the management of weeds and crop pathogens. The complexity of agroecosystem functioning suggests recognising that tillage is not inherently at odds with agroecological principles. When deployed strategically, it can support ecological intensification by reducing pesticide dependency. Rather than rejecting tillage outright, agroecological transitions should reclaim it as a context-specific, service-providing practice that complements other management measures, like crop rotation, cover crops, and varietal diversity.

Soil carbon and tillage: beyond the surface debate

Reduced tillage has become emblematic of climate-smart agriculture, largely due to its association with increased SOC concentration in the surface layer. However, this observation has frequently been overinterpreted as evidence of superior SOC sequestration, while an in-depth assessment reveals a more nuanced picture^34,35. Indeed, no-till first results in a redistribution of SOC with a strong vertical stratification, concentrating SOC near the surface while potentially depleting deeper layers³⁶. Secondly, a change in tillage alters soil bulk density: it is generally increased under no-till compared to conventional tillage across much of the soil profile³⁴. Therefore, SOC stocks of different tillage practices must be compared at a sufficient depth, using equivalent soil mass approaches³⁷.

Over the last decade, several meta-analyses have been published that compare SOC stocks over equivalent soil masses down to 30 cm or more, between continuous no-till (or reduced tillage) and conventional tillage^{36,38,39,40,41}. These meta-analyses, which gathered experimental data from various geographical contexts (Europe, China, boreo-temperate regions and the world), showed a moderate increase in SOC stock compared to conventional tillage for the 0–30 cm soil layer, of 0.23 t C ha⁻¹ yr⁻¹ on average (Table 2). This was even reduced by considering deeper soil layers, which led to non-significant SOC storage in two out of three studies. Experimental data were also highly variable (mean s.d. = 0.62 t C ha⁻¹ yr⁻¹ for the 0–30 cm soil layer), probably due to pre-existing spatial variability as most data came from synchronic measurements⁴². Long-term diachronic (time series) studies are rare in the scientific literature, but essential for a better understanding of the impact of tillage on SOC. In northern France, the closely-monitored long-term tillage experiment of Boigneville showed no significant differences in total SOC stocks between no-till, shallow tillage and conventional tillage after 47 years⁴³.

Table 2 Results from five meta-analyses considering tillage effect on SOC stocks at equivalent soil mass in the layer 0–30 cm or deeper

These results challenge the hypothesis that SOC is more physically protected in no-till systems, reducing its turnover due to lower contact between microbial decomposers and organic substrates. While this hypothesis of a lower turnover is included in some agronomic or biogeochemical models, the magnitude of the effect varies widely between them⁴⁴. Some studies have used ¹³C natural abundance techniques to study the SOC dynamics in relation to tillage systems when there has been a change in vegetation from C3 to C4, or vice versa. While Six et al.⁴⁵ found a higher SOC mean residence time (i.e. a lower SOC turnover) under no-till compared to conventional tillage, three other studies found no significant differences between tillage treatments^43,46,47. In particular, Haile-Mariam et al.⁴⁷ found similar mean residence times between conventional tillage and no-till for different SOC fractions in three long-term experiments in the USA under continuous maize cultivation.

Although tillage can stimulate the mineralisation of soil organic matter in the short term (e.g. ref. ⁴⁸), ceasing no-till practices by adopting occasional or regular tillage has minimal effects on SOC stocks. We synthesised experimental studies that monitored SOC stocks after one or several tillage operations following several years of continuous no-till ( ≥ 6 years). Of the twenty-one situations examined (eight from Canada, five from the USA, five from Australia, one from Argentina, one from China and one from France), only two showed a significant decrease in SOC stocks (Table 3). The main effect of an occasional tillage in long-term no-till systems is reducing SOC stratification⁴⁹.

Table 3 Impacts on soil organic carbon stocks of one or several tillage operations following several years of continuous no-till: a synthesis of field-based studies

Taken together, these findings suggest that the debate around tillage and SOC sequestration has often been framed too simplistically. Tillage affects the distribution of SOC more than its total stock. Its influence must therefore be evaluated in conjunction with other factors, particularly net carbon inputs as influenced by cropping system management, soil mineralogy and aggregation, and climate. In short, tillage reduction is not a guaranteed pathway to enhanced SOC storage.

Carbon inputs drive soil carbon sequestration more than tillage

While tillage mixes the soil and influences the spatial accessibility of organic materials for decomposers, resulting in the mineralisation of a part of SOC, long-term changes in SOC stocks are overwhelmingly governed by the quantity and diversity of carbon inputs at different depths. Indeed, the microbial carbon assimilation process of these inputs may form microbial compounds which persist in the soil by incorporation into aggregates and adsorption on soil surfaces, counterbalancing the transient higher microbial respiration induced by soil mixing^50,51. Evidence from field experiments points to above- and below-ground biomass inputs as the primary drivers of SOC sequestration at the ecosystem scale, thereby assigning relative importance among small-scale processes that regulate carbon persistence in soil. Consequently, systems with higher net primary productivity, due to multi-species cover crops, pluriannual legumes, diversified crop rotations with deep-rooted or perennial species, the integration of forage or ley phases, the retention of crop residues, or agroforestry, tend to accumulate more SOC, regardless of the tillage regime^41,52,53. In practical agroecological terms, this SOC sequestration is promoted by management practices repeated over time and space that increase plant productivity and diversity, resulting in increasing the carbon inputs from plant residues returned to soil, extend the period of living roots, and enhance the organic amendments. These practices directly support the processes depicted in Fig. 1, notably increased the quantity and molecular diversity of carbon inputs, enhanced soil life, and physical protection by aggregate formation or organo-mineral association of SOC.

This insight is supported by long-term cropping system experiments. A 25-year experiment at the Kellogg Biological Station in Michigan (USA) showed that systems with tillage and cover crops stored twice as much SOC as a no-till system without cover crops (0.4–0.5 vs 0.2 t C ha⁻¹ yr⁻¹, respectively)⁵⁴. In the 16-year “La Cage” experiment in Versailles (France), conservation agriculture with no-till and a permanent cover crop increased SOC stocks by 0.63 t C ha⁻¹ yr⁻¹, compared to only 0.08 t C ha⁻¹ yr⁻¹ in the conventional system⁵⁵. Modelling with the AMG SOC model revealed that these gains could be fully explained by increased crop C inputs, rather than changes in mineralisation rates due to tillage⁵⁵. Soil incubations confirmed similar specific SOC mineralisation rates between systems⁵⁶. Similarly, in two long-term experiments in subtropical climates that combined tillage and crop rotation or residue management treatments (one in Brazil and one in Mexico), the differences in SOC stocks were mainly explained by the differences in crop C inputs, rather than by tillage effects^57,58.

Particularly relevant is the role of root-derived carbon. Root biomass is more efficiently stabilised in the soil matrix compared to above-ground residues^59,60. This aligns with the 'root-centric' strategy for SOC sequestration^61,62, emphasising that increasing root inputs is a more reliable pathway for building long-lasting soil carbon than simply reducing mechanical disturbance. Cover crops play a pivotal role in this context⁶³. Meta-analyses have shown that their use, especially where sufficient cover crop biomass is generated⁶⁴, increases SOC by 0.2–0.9 t C ha⁻¹ yr⁻¹ on average, while also improving aggregate stability and microbial activity^52,65,66. Furthermore, legume-based cover crops also provide a nitrogen green manure service⁶⁷.

This reorients the debate: rather than asking “how much tillage is acceptable?”, we should ask “how can we enhance and stabilise organic inputs within resilient agroecosystems?” From an agroecological perspective, this implies combining practices such as diversified rotations, permanent or temporary soil cover, increased functional diversity of crops, and reduced reliance on external inputs. Importantly, many of these practices also contribute to non-chemical weed and disease regulation, offering concrete technical pathways for designing cropping systems that both reduce pesticide use and enhance climate change mitigation potential.

GHG trade-offs: is N₂O the hidden cost of no-till?

While no-till practices are often promoted for their carbon benefits, they can introduce unintended consequences for other GHG emissions, particularly nitrous oxide (N₂O), whose potential to exacerbate global warming is 273 times greater than that of CO₂⁶⁸. As cropping systems reduce soil tillage and increasingly rely on surface residues and cover crops, conditions may favour the main microbial processes responsible for N₂O production, nitrification and denitrification⁶⁹. Although emissions are mainly linked to N fertilisation⁷⁰, they can also arise from organic matter transformation processes. Meta-analyses across diverse pedoclimatic zones and cropping systems showed that no-till and reduced tillage have inconsistent effects on N₂O emissions (Fig. 3), with either slightly lower, equal, or slightly higher emissions compared to conventional tillage^{71,72,73,74,75,76}, possibly offsetting the benefits of SOC storage⁷⁷. The magnitude and direction of this effect depend on soil texture, climate, and residue management. Fine-textured, poorly aerated soils in humid climates are especially prone to increased N₂O emissions under no-till, where surface mulches maintain soil moisture and limit oxygen diffusion, creating ideal conditions for denitrification hotspots^72,73,76. The aforementioned “La Cage” field experiment exemplifies this pattern: although conservation agriculture showed high levels of SOC storage, its N₂O emissions were 1.7 and 3.4 times higher than those of conventional and low-input systems, respectively⁷⁸. System maturity could also influence the response of N₂O emissions to no-till. Several meta-analyses have shown that no-till duration had a significant effect on relative N₂O emissions compared to conventional tillage. However, while three meta-analyses^71,73,75 showed that relative N₂O emissions decreased with no-till duration, the opposite trend was observed by Li et al.⁷⁴.

Fig. 3: Mean effect of no-till (NT) and reduced tillage (RT) on N₂O emissions relative to conventional tillage, according to six meta-analyses.

Error bars are the 95% confidence intervals. Numbers in parentheses indicate the number of comparisons. These meta-analyses used paired comparisons between NT or RT and conventional tillage, with N₂O emissions measured in the field for the whole crop season. The studies were not geographically limited, except for that of Hashimi et al.⁷⁶, which focused on drylands (arid and semi-arid climates). However, the meta-analyses had different focuses and therefore different inclusion/exclusion criteria. For example, van Kessel et al.⁷⁵ excluded flooded systems such as rice paddies. Feng et al.⁷¹ only included studies that measured both N₂O and CH₄, with a substantial proportion (~ 36%) of paddy rice experiments in the dataset. Huang et al.⁷² and Li et al.⁷⁴ focused on cereal crops: barley, maize, rice, wheat, and also soybean in the case of Li et al.⁷⁴.

These findings highlight the need for integrated GHG assessments that consider both SOC and N₂O dynamics, as well as fossil CO₂ emissions⁷⁹. Although no-till reduces CO₂ emissions due to fuel combustion, this reduction is generally low compared to the GHG emissions from other agricultural inputs, particularly fertilisers^78,80. Given the contrasting effects on N₂O emissions, blindly promoting no-till as the key practice for climate mitigation, without considering cropping system diversity and tillage intensity, residue management, and local soil conditions, risks delivering ambiguous or even counterproductive outcomes.

Conclusions and perspectives

The debate concerning the convergence of challenges related to pesticide reduction and climate-smart agriculture is often oversimplified, with the role of tillage reduced to a binary narrative, where no-till is presented as the solution, and all other tillage practices, irrespective of intensity of soil disturbance, as the problem. Yet, evidence from long-term experiments, meta-analyses, and farm-level observations reveals a more complex reality. Tillage, when used strategically, can provide crucial ecosystem services, particularly for non-chemical management of cover crops and weeds, pests, and soil-borne pathogens, without necessarily undermining SOC stocks or GHG goals (Fig. 2).

Our synthesis highlights three key points. First, tillage remains an essential tool for pest and weed control in low-input or organic systems. Second, long-term SOC sequestration is an emergent system property, governed more by net system carbon inputs, soil biological life and climate than by tillage. Third, reduced or no tillage can increase N₂O emissions under certain conditions, calling for a whole-system approach to reduce GHG emissions.

It is important to note, however, that tillage also has other significant impacts not fully addressed in this discussion. For instance, soil disturbance strongly influences erosion dynamics, potentially accelerating soil degradation in certain contexts. Moreover, repeated tillage affects soil biological activity, including macrofauna such as earthworms and the structure of mycorrhizal fungal communities, which play key roles in nutrient cycling, soil aggregation, and plant health. Considering these dimensions could lead to a different interpretation of the trade-offs discussed here and highlights the need for a broader perspective on the ecological consequences of tillage.

Rather than advocating for universal tillage or no-till adoption, we argue for a more pragmatic and context-sensitive framework. This requires recognising tillage as a situational lever within agroecological design, to be mobilised or constrained based on system goals, local constraints, and trade-offs across productivity, soil health, biodiversity, and climate outcomes.

Future research should integrate process-based modelling with empirical data to simulate long-term interactions among SOC dynamics, microbial activity, and GHG emissions. The development of indicators such as the microbial carbon use efficiency (CUE) also offers new pathways to monitor and optimise carbon retention in diverse agricultural systems.

Ultimately, designing resilient, climate-positive cropping systems with low pesticide inputs will require more than banning tillage: it will demand managing complexity through adaptive, systems-based thinking.