It has been well-established that tilling the soil leads to losses of SOC as cultivation breaks up the soil and exposes organic matter previously protected within soil macroaggregates to microbial decay (Beare et al., 1994; Six et al., 2000). Cultivation also incorporates and fragments plant material, increasing its vulnerability to microbial attack (Hendrix et al., 1986; Beare et al., 1993). Decreasing the amount of tillage or introducing no-till (NT) thus has the potential to decrease the amount of SOC lost from the profile by decreasing the turnover rate of macroaggregates, increasing the physical protection of particulate organic material, and reducing soil to residue contact. For example, in a Brazilian Acrisol under cereal cropping, turnover times of 17 vs. 36 years were observed in conventional and CA, respectively, due to the reduced disturbance in the NT CA system (Bayer et al., 2006a).

However, it should be noted that in some instances tillage has also been observed to increase SOC stores relative to CA systems (Chan et al., 1992; Christopher et al., 2009; Blanco-Canqui et al., 2011; Zhang et al., 2014). In areas where full inversion tillage is carried out, residues may be buried in a region where poor soil aeration can limit decomposition (relative to the soil surface), particularly under cool, moist climatic conditions (Gregorich et al., 2005; Christopher et al., 2009).

Crop residues can be defined as biomass remaining on the soil's surface after harvest. In some systems, linear increases in SOC stocks can be observed with increasing rates of residue addition (Duiker and Lal, 2002; Virto et al., 2012; Liu et al., 2014). Thus, in CA systems, which emphasize the retention of residues on the soil surface, greater residue input can potentially lead to greater SOC storage. However, where residue production is low e.g., due to low soil fertility or the presence of soil constraints, there can be insufficient residue retention under CA to positively impact SOC stores (Virto et al., 2012; Palm et al., 2014). Indeed, in some situations CA can even lead to a reduction in plant growth (e.g., due to lower soil temperatures, compaction or greater disease/weed/pest pressure) and this can also lead to decreases in residue inputs and lower or no SOC sequestration (Yang et al., 2013). In areas with low fertility, integrated nutrient management is essential to ensure a buildup of SOC and the success of CA systems. Low fertility results in low crop yields, which leads to reduced organic matter input and hence, lower SOC (Lal, 2015b). This cycle can only be broken by the addition of nutrients to the soil system via external fertilizer input (organic or inorganic) and/or the incorporation of legumes into cropping rotations.

Different crops may have different effects on the quantity, quality, and periodicity of C inputs and can modify the soil in different ways (e.g., rates of water extraction, nutrient use), which can influence mineralisation rates and the growth of subsequent crops (Huggins et al., 2007). Thus, differences in crop rotation between CA and conventional agricultural systems also have the potential to impact SOC values. The elimination of monocultures and incorporation of plant species into rotations that return greater amounts of residue to the soil are often associated with greater SOC stock in CA systems (Huggins et al., 2007; dos Santos et al., 2011; Conceição et al., 2013). In some systems, root input, in particular, has been found to be important. For example, in one study conducted on a Brazillian Ferrosol of the long-term (17 years) contribution of cover crop or forage based NT rotations to SOC stocks, SOC stocks showed a close relationship with the root additions of different plant species (dos Santos et al., 2011). Where legumes are included, these can also add additional N to the system, which can enhance soil fertility and subsequent crop biomass production (Hansen et al., 2012; Mbuthia et al., 2015; Raphael et al., 2016; Veloso et al., 2018). The maintenance of residue cover can also decrease processes that can limit the growth and biomass production of the main crop, such as erosion, nutrient leaching, and weeds, pests and diseases (Tittonell et al., 2012; Gabriel et al., 2013; Veloso et al., 2018; Williams et al., 2018).

The exact magnitude of the difference in SOC between CA and conventional agricultural systems varies greatly and is influenced by many factors, including climate (Ogle et al., 2005, 2019), soil type (Liang et al., 2002; Nyamangara et al., 2014), baseline SOC (VandenBygaart et al., 2003; Steinbach and Alvarez, 2006), crop management (VandenBygaart et al., 2003), time since management change (González-Sánchez et al., 2012), and sampling depth and methodology (Olson, 2013). This means that estimates of the magnitude of the change following conversion to CA vary widely (Table 1). For example, estimates ranging from −0.15 Mg/ha/year in areas such as the midwestern USA (Christopher et al., 2009) to +0.93 Mg/ha/years in tropical Brazil (Bernoux et al., 2006) have been recorded. However, it is clear that in regions where soil and climatic conditions are favorable for biomass production and where negative yield impacts are not observed, then CA systems will often have higher amounts of SOC relative to conventionally managed systems, particularly in the surface of the soil profile. Although, the large range in the sequestration rates observed indicates that the magnitude of change is likely to be highly site specific. It should also be noted that most of the literature regarding the effect of CA on soil properties has focused on NT and residue retention v conventional tillage and residue removal. Fewer studies have also considered the impact of species diversification and the inclusion of practices such as cover crops and intercropping on SOC change. Where these are carried out they are often found to lead to greater increases in SOC than NT and stubble retention alone, and the effect of management changes on soil properties would thus often be greater when all elements of the CA system, which may have interactive and synergic effects, are incorporated together (Blanco-Canqui and Ruis, 2018).

Where improvements in soil properties are observed under a system of CA management, there are many instances where this has been associated with yield increases (Table 4). However, no change and decreases in yield have also been observed (Table 4). A recent global meta-analysis of the effect of CA practices on yield concluded that CA (NT+residue retention+crop diversification) on average worldwide leads to yield reductions of 2.5% (Pittelkow et al., 2015). Where only partial implementation of CA occurs greater yield loss can be expected (e.g., −9.9% for NT alone, −5.2% for NT+residue retention). However, yield increases were observed under certain circumstances, for example, gains of 7.3% were observed under rainfed agriculture in dry climates, and the yield declines observed overall tended to decrease the longer CA had been implemented.

The yield gains frequently observed in dry climates are commonly attributed to the improvements in soil water storage under CA (Halvorson et al., 2000; Hemmat and Eskandari, 2004, 2006; Li et al., 2007; Radford and and Thornton, 2011; Araya et al., 2012; Soane et al., 2012; Steward et al., 2018; Garcia-Palacios et al., 2019; Page et al., 2019). The possibility for early planting is also greater in CA systems due to the reduced land preparation requirements, which may result in more efficient use of rainfall and reduced risk of crop failure due to terminal heat stress (Kumar et al., 2018; Devkota et al., 2019). In hot regions, the lower soil temperatures under CA systems due to residue retention can also help improve plant growth and yield (Zhang et al., 2014; Steward et al., 2018). Improvements in soil fertility and/or reductions in pests/diseases as a result of the inclusion of legumes and diversified species into cropping rotations have also been reported to contribute to yield increases in some instances (Rusinamhodzi et al., 2011; Ngwira et al., 2012; Nyagumbo et al., 2016; Kaye and Quemada, 2017).

Where yield decreases are observed, these are commonly associated with greater waterlogging in response to the greater water infiltration and lower evaporation in CA systems—with higher rainfall areas and poorly drained soils the most significantly affected (Rusinamhodzi et al., 2011; Nyagumbo et al., 2016; Lal, 2018; Steward et al., 2018). Decreased soil temperatures can also delay plant maturity, particularly in cooler, high latitude regions, and negatively affect yield (Hatfield et al., 2001; Wang et al., 2007; Rusinamhodzi et al., 2011; Soane et al., 2012; Zhang et al., 2014; Lal, 2015b). Consequently, in cold, wet environments with poorly drained soils, CA is often unlikely to lead to yield increases. Other instances of yield loss have been associated with:

Inadequate control of weeds, pests and/or diseases in the absence of tillage (Soane et al., 2012; Dang et al., 2015; Li et al., 2019a);

Soil structural issues, such as compaction or surface sealing, due to the absence of tillage and inadequate residue cover/biomass production, particularly in drier regions with infertile soils (Rusinamhodzi et al., 2011; Baudron et al., 2012; Kienzler et al., 2012; Vanlauwe et al., 2014; Giller et al., 2015; Thierfelder et al., 2015); and

Lack of access to effective seeding equipment, required inputs (herbicides, fertilizers), and knowledge regarding how best to manage CA systems (nutrition, weeds, pest and diseases) (Rusinamhodzi et al., 2011; Kienzler et al., 2012; Thierfelder et al., 2015).

The poor yields observed due to these reasons can often be improved by better customizing the CA system to a particular region or by removing fertility constraints to plant growth to improve residue cover. The yield loss observed in the early years of CA adoption before improvements in SOC and associated structural/fertility improvements occur may also often reverse over time as soil properties and nutrient availability improve and farmers/researchers gain greater skills in managing the CA system (Araya et al., 2012; Soane et al., 2012; Brouder and Gomez-Macpherson, 2014; Pittelkow et al., 2015; Vastola et al., 2017; Steward et al., 2018).

It is often stated that the CA system can help agriculture adapt to increasing climate variability and the occurrence of extreme events (Kassam et al., 2009; Williams et al., 2018). The erosion protection, reduction in soil temperatures and improved infiltration rates, for example, can help deal with more intense rainstorms, and increased daily temperature ranges and frequency of drought (Kassam et al., 2009). In particular, the yield advantage often observed in CA in dry years is likely to provide an advantage in environments where rainfall is, or is becoming, more variable (Vastola et al., 2017). In agreement with this, some analyses have found that CA practices, and in particular the incorporation of cover crops, are associated with increased crop yield stability (Williams et al., 2018; Page et al., 2019; Anderson et al., 2020). Although, other analyses have found that there is no difference in the yield stability of CA vs. conventional agricultural systems (Rusinamhodzi et al., 2011; Daigh et al., 2018; Knapp and van der Heijden, 2018).

Where CA leads to similar or greater yields, profitability is generally improved due to reduced costs of land preparation and labor, and reduced water requirements (in irrigated systems) (Ngwira et al., 2012; Jat et al., 2014; Kumar et al., 2018; Devkota et al., 2019). Savings in fertilizer use, and in some instances reduced herbicide and insecticide costs, can also increase profitability (Kassam et al., 2009; LaCanne and Lundgren, 2018). Indeed, in some instances, even when yield reductions are observed, CA systems can still be more profitable than conventional agricultural systems due to reduced input costs (Vastola et al., 2017; LaCanne and Lundgren, 2018). Examples of improvements in the profitability of CA vs. conventional agricultural systems can be seen in Table 5.

However, increases in profitability are not a universally observed trend. Where markets for alternative crop rotations are absent, or offer a reduced price compared to the main cash crop, losses in net farm income can be observed despite input cost savings (Giller et al., 2015; TerAvest et al., 2019). Indeed, lack of access to seed, suitable markets, transportation facilities and/or suitable storage facilities can sometimes be a significant barrier to the profitable use of diversified crop rotations, which are often an important part of the success of CA systems (Andersson and D'Souza, 2014; Ares et al., 2015; Carlisle, 2016; Brown et al., 2017; Dhar et al., 2018). In small holder operations some input costs, such as labor requirements, may also not be monetized, making the cost reductions that can be realized with CA small (Giller et al., 2015). Indeed, particularly in the early stages of CA system adoption before soil fertility increases and weeds, pests and diseases are effectively controlled, input costs may actually increase as new tools (e.g., planting equipment), additional fertilizers, and herbicides are required (Andersson and D'Souza, 2014; Dhar et al., 2018). In smallholder operations, CA can also often be associated with greater labor requirements in instances where hand weeding is required due to the lack of herbicide availability (Giller et al., 2015). Opportunity costs can also be significant, particularly for residue use when these can also be profitably used as stock feed, or baled for sale to other producers (Andersson and D'Souza, 2014; Ares et al., 2015; Beuchelt et al., 2015).

The successful control of weeds, pests and diseases in the absence of tillage, the stratification of immobile nutrients at the surface of the profile, poor plant establishment due to inadequate seeding equipment, decreased N availability, and the development of soil structural issues, such as surface crust and compaction, can all be significant constraints to crop production in CA systems (Dang et al., 2015). In addition, in areas where low yields are the norm, it can be difficult to build up SOC and residue cover to improve soil fertility, prevent erosion and suppress weeds (Andersson and D'Souza, 2014.). In many instances these agronomic challenges can be overcome through better adaptation of the principles of CA to local conditions. For example:

The development of locally appropriate seeding equipment and the identification of locally adapted crop rotations/cover crops can help deal with the problems of plant establishment and weed, pest and disease pressures in the absence of tillage (Verhulst et al., 2010; Thierfelder et al., 2018).

Appropriate application of fertilizers and/or better incorporation of legumes into cropping systems can help deal with problems of N availability and allow greater plant biomass production to increase residue cover and improve soil fertility (Verhulst et al., 2010).

The incorporation of cover crops can maintain soil cover during periods when the main crop is not growing and add additional organic material to the soil where crop residue production is low (Fageria et al., 2005; Tittonell et al., 2012; Veloso et al., 2018; Williams et al., 2018).

The use of strategic tillage, which is the practice of occasionally cultivating NT soils, can help with weed, pest and disease management, nutrient stratification, and compaction in continuous long-term NT systems (Kirkegaard et al., 2014; Dang et al., 2015, 2018).

However, in order to successfully identify practices appropriate for different geographical regions and cropping systems, adequate research is required to develop effective and locally adapted approaches. In particular, the identification of locally appropriate crop rotations is important, as is greater research into strategies such as strategic tillage to understand the pros and cons of this type of approach and optimum frequency and timing of tillage for a range of soils and agro-climatic regions.

The successful implementation of CA systems must also take into account not just technical and agronomic issues, but social, cultural and institutional factors (Lal, 2015b). For example, in many farming communities, some of the principles of CA agriculture, particularly the use of NT, may run counter to many established land management traditions that have worked for generations and which have often created cultural values and rural traditions (Gonzalez-Sanchez et al., 2016; Tekle, 2016). Practices such as NT may be viewed as a “lazy” and ineffective way of farming, and farmers who initially adopt CA can risk backlash from their community (Zhang et al., 2014; Tekle, 2016). In some situations, it can even foster social conflict due to competing demands for crop residues (e.g., from stock grazing) (Knapp and van der Heijden, 2018). Overcoming the mindset in farming communities that tillage is required for successful agricultural production and addressing issues around residue use is thus required to increase the uptake of CA systems (Bhan and Behera, 2014; Gonzalez-Sanchez et al., 2016; Tekle, 2016).

Similarly, farmer access to information regarding the benefits of CA and how to implement it effectively is an important precursor for adoption and continued uptake over time (Llewellyn et al., 2012; Abdulai, 2016; Carlisle, 2016). Indeed, the complexity of CA systems and the need to understand how to use agronomic techniques to manage weed/ pest/ diseases and ensure sufficient crop nutrient availability mean that the implementation of CA without sufficient education around the management of its challenges is unlikely to result in long-term success. Effective extension and the development of partnerships between farmers, researchers and the private sector (e.g., seed companies, machinery suppliers) to refine CA systems and demonstrate their benefits is required to facilitate knowledge dissemination, increase CA uptake and reducing dis-adoption rates over time (Dauphin, 2003; Bellotti and Rochecouste, 2014; Bhan and Behera, 2014; de Freitas and Landers, 2014; Ares et al., 2015).

Conservation agriculture also needs to be demonstrated to farmers as a profitable way of farming before widespread adoption will occur. This may require a degree of structural change to address issues such as the lack of markets for produce (particularly for diversified crop rotations) and the inability to economically obtain required inputs such as seed, fertilizer, herbicide, and locally adapted machinery (Corbeels et al., 2014; Carlisle, 2016; Tekle, 2016; Brown et al., 2017). The adequate availability of fertilizers can be particularly important in areas with inherently low soil fertility where the production of sufficient quantities of crop residue for soil protection is difficult (Giller et al., 2015; Thierfelder et al., 2018). The presence of protectionist mechanisms, such as subsidies for the production of certain commodities, can also hinder adoption as these do not encourage the reduction of production costs and can act to decrease the profitability of CA practices (Carvalho and Lourenco, 2014; Tekle, 2016).

Additional assistance may also be required for poorer farmers who are less likely to adopt CA due to the initial investment required around establishment and the risk associated with decreased yields early in the adoption process (Abdulai, 2016; Brown et al., 2017; Ding, 2018; Harper et al., 2018). Policy initiatives to subsidize or incentivize NT, for example by increasing access to credit and the farm machinery required for NT operations (e.g., rental schemes, custom hire) can increase adoption (Bhan and Behera, 2014; Abdulai, 2016; Carlisle, 2016). However, ongoing support and the removal of larger institutional barriers to adoption are essential for farmers to continue with CA once initial incentives have been removed (Andersson and D'Souza, 2014; Brown et al., 2017; Chinse et al., 2019). Indeed, until institutional and/or social barriers are able to be overcome, CA is unlikely to be a successful system of sustainable intensification.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.