How do we cultivate in England? Tillage practices in crop production systems

Abstract Reducing tillage intensity offers the possibility of moving towards sustainable intensification objectives. Reduced tillage (RT) practices, where the plough is not used, can provide a number of environmental and financial benefits, particularly for soil erosion control. Based on 2010 harvest year data from the nationally stratified Farm Business Survey and drawing on a sub‐sample of 249 English arable farmers, we estimate that approximately 32% of arable land was established under RT, with 46% of farms using some form of RT. Farms more likely to use some form of RT were larger, located in the East Midlands and South East of England and classified as ‘Cereals’ farms. Application of RT techniques was not determined by the age or education level of the farmer. Individual crops impacted the choice of land preparation, with wheat and oilseed rape being more frequently planted after RT than field beans and root crops, which were almost always planted after ploughing. This result suggests there can be limitations to the applicability of RT. Average tillage depth was only slightly shallower for RT practices than ploughing, suggesting that the predominant RT practices are quite demanding in their energy use. Policy makers seeking to increase sustainable RT uptake will need to address farm‐level capital investment constraints and target policies on farms growing crops, such as wheat and oilseed rape, that are better suited to RT practices.


Introduction
Lowering tillage intensity is a management practice that potentially can reduce environmental impacts and improve agricultural outputs (i.e. providing 'sustainable intensification'; Buckwell et al., 2014). A number of tillage systems forego the use of the plough (i.e. do not involve soil inversion). These systems vary in the extent to which soil is disturbed, ranging from extensive, in deep reduced tillage (RT 1 ), to limited, in shallow RT and minor, in zero-tillage (ZT) with the latter establishing the crop with only very minimum soil disturbance (Davies & Finney, 2002). These tillage systems are often part of wider agronomic practices that include residue management and, in conservation agriculture, continuous groundcover and diverse crop rotations. The definition of tillage systems varies among practitioners; Table 1 provides an overview summary of these typical tillage/establishment systems. In this paper 'reduced tillage' is used to refer to cultivation systems that do not involve soil inversion.
Previous authors have identified environmental benefits of reducing tillage intensity in terms of: reduced soil erosion, pesticide runoff, nitrate leaching and watercourse sedimentation; improved soil quality and smaller greenhouse gas emissions (Fawcett & Towery, 2002;Holland, 2004;Morris et al., 2010). Other benefits include reduced fuel costs, improved timeliness of field preparation, less machinery input required and lower machinery expenses through lower wear and tear (Baker et al., 2007;SoCo Project Team, 2009). Across a range of studies conducted in northern Europe, labour requirements were 30-40% lower for RT systems relative to ploughing (SoCo Project Team, 2009); fuel use was also much lower, although this was greatly dependent on specific machinery used and the soil conditions. Throughout the literature, there is inconsistency in the impacts of adopting RT practices and this is due to the variation in RT practices used as well as the specific cropping systems, soil types and climate.
Reduced tillage is widely practised in North and South America, with increasing uptake in South Africa, Australia and other semi-arid areas where the primary driver for uptake is reduction in soil erosion (Holland, 2004). In humid temperate regions, such as Northwest Europe, where soil erosion is less of a problem, the main reason for the use of RT is cost savings (Morris et al., 2010). Davies & Finney (2002) reviewed the use of RT in the UK and suggested that using RT practices offered the best opportunity for reducing labour and machinery costs and, therefore, reducing production costs and increasing profit.
Estimates for the UK suggest that one-third of crops grown on cereal-growing land were established under RT in the 1970s; however, problems with grass weeds led to a reduction in use to about 10% in 1988 (Davies & Finney, 2002). During that time, straw burning facilitated the use of RT because residues left in situ can interfere with seed drilling and result in increased disease and pest incidence (Carter, 1994). In the absence of straw burning, these issues have been previously cited as constraints to the use of RT (Cannell, 1985). The phasing out of straw burning, which culminated in a ban on the practice in 1993 for England and Wales, further encouraged use of the plough. More recently, RT use has been increasing: Defra's 2010 Farm Practices Survey found that 44% of arable land in England was under RT (with at least 30% of residue coverage) with 4% of this under ZT (Defra, 2010). Data from the 2010 Survey on Agricultural Production Methods (SAPM) show that tillage practices vary throughout Europe (EuroStat, 2013). RT in the UK 2 was estimated as 39%, which is similar to Germany (41%) and France (36%), yet much greater than the Europewide average of 26%. ZT covered 5% of arable land in the UK whilst Germany and France had 1% and 4%, respectively. Although these national usage figures exist, there is limited data available on specific tillage practices at the farm level. However, interest in adoption of RT may be increasing as it is recognized as an important management practice and the Soil Protection Review, which must be completed by every farmer receiving payments as part of the Common Agricultural Policy in the UK, encourages the use of RT practices (Anon, 2010). In addition, policy makers are seeking to incentivize sustainable intensification practices to achieve greater levels of production from lower input use (e.g. via the UK funded Agri-Tech programme). RT is recognized as a practice contributing to sustainable intensification. Complicating comparisons between information sources is that some statistics are given for the UK as a whole whilst others refer to individual countries within the UK. The data from the survey presented in this paper refer to England.
Although the area under RT appears to be increasing, there are constraints to its wider adoption (Powlson et al., 2012). The feasibility of adopting RT depends on soil type and climate (Cannell et al., 1978;Carter, 1994;Davies & Finney, 2002;Morris et al., 2010). In particular, good drainage and a naturally stable structure are desirable to avoid soil compaction. Calcareous soils with or without large clay contents are more suited to RT (Davies & Finney, 2002) than light sandy soils (Morris et al., 2010). Where ZT is practised in the UK, it tends to be on calcareous clay soils, which 'self-mulch' (Powlson et al., 2012). It can be assumed that these constraints to adoption are likely to be greater the lower the level of tillage intensity.
Reduced tillage systems in much of Northwest Europe tend to require greater herbicide use to control weeds (Melander et al., 2013). Even with additional herbicide use, weed problems can limit the use of RT. In particular, blackgrass (Alopecurus myosuroides Huds.) is the major weed of winter cereals and one method of control is mouldboard ploughing (see, e.g. Lutman et al., 2013). Due to this, a common approach in the UK is to use RT practices in conjunction with ploughing; Powlson et al. (2012) refer to this as rotational ploughing, whereby the plough is used every 3 to 4 yrs. There is not, however, a strict definition of rotational ploughing, and it is likely that farms range from using ploughing at a fixed point within a set rotation to using ploughing in response to specific conditions (e.g. after outbreaks of disease or pests). In Table 1 these are broadly defined as rotational ploughing and strategic tillage, respectively, with the term mixed tillage system referring to farm systems utilizing both ploughing and RT.
Although there has been considerable work investigating the impact of continuous ZT on cropping systems and environmental impacts relative to yearly ploughing (see Soane et al., 2012 for a review), work quantifying changes resulting from other RT systems and mixed tillage systems compared to yearly ploughing is limited and results are often confounded by location and specific cultivation practice outcomes. Dang et al. (2015a,b) reviewed the literature on using strategic tillage in ZT systems and found that impacts on cropping systems and environmental impacts were highly variable.
Evidence from field experiments comparing different tillage systems shows a slight reduction in crop yields under RT ( Van den Putte et al., 2010;Arvidsson et al., 2014); however, yield impacts varied with crop type, tillage depth and crop rotation. When individual field experiments are considered, yields can be higher under RT (e.g. Knight, 2004;Verch et al., 2009). Yield data for different tillage systems in the UK are, however, limited and moreover typically relate to experiments that were undertaken prior to the ban on straw burning in England and Wales discussed above. As straw burning reduces weed, disease and pest burden for the following crop (Graham et al., 1986), these studies are unlikely to reflect current evidence on the impact of RT on crop yields. Alongside this, there have been advances in RT machinery, improving the establishment of crops in RT systems. A potential issue with these studies is that crops are drilled simultaneously on ploughed and RT experimental plots; in commercial contexts RT practices tend to be quicker to implement than ploughing allowing earlier crop establishment, which could provide a yield benefit not captured in these studies.
Given this background, the aim of this paper is to present a more current analysis of the uptake of RT systems on English arable farms, providing definitive context to the tillage depth employed within commercial farm practice, the extent of continuous or rotational RT and yield and crop gross margins associated with different tillage systems. Specifically we test the following hypotheses: • Area of RT is influenced by location, farm type, farmer age and level of education, and farm size.
• Average depth of main type of tillage is influenced by location, farm type, farmer age and level of education, and farm size.
• Choice of tillage depends on crop type. • Tillage practices influence production metrics.

Methodology
Data were collected between February and October 2011 in conjunction with a survey of English arable farms, sampled from the Farm Business Survey (FBS) research programme which samples approximately 3% of the commercial farm business in England. The information was obtained at the same time as that reported by Glithero et al. (2013a,b,c) on the supply of biomass for biofuel production. A detailed description of the data collection method is provided in Glithero et al. (2013c). Analyses in the papers cited above were based on a sample of 249 farms businesses, drawn from the nationally stratified FBS to form a stratified subsample of commercial farm businesses within farm types of interest. Specifically, this subsample represented approximately 46% of FBS farms within the three main arable farm types (Cereals, General Cropping and Mixed). Within the FBS, farm types were classified on the basis of economic output from the enterprises or group of enterprises. When a farm business derived at least 2/3rd of its output from an individual enterprise or grouping of enterprises (e.g. combinable crops output for the Cereals farm type) the farm business was classified as a particular farm type; in the absence of the 2/3rd output threshold being met from a specific enterprise, farms were classified as Mixed. Additionally farms were stratified across three utilized agricultural area groupings within these farm types (Anon, n.d.). These 249 farms covered approximately 1.5% of the total commercial population of these farm types in England. Corresponding individual farm data for the main crop specific gross margins (GMs, i.e. value of total output less variable costs), yields and variable costs were taken from the FBS data for the 2010/11 financial year; the FBS did not record detailed records of residue use (e.g. straw incorporation versus baling); however, data on residue use in England were presented in Glithero et al. (2013c) drawing on the survey detailed above. The economic data were provided by the main FBS account for each of the 249 farm businesses. Here, we focus on the financial records at the level of the total output and variable costs for individual crop enterprise without the data relating to the fixed costs of production. Drawing on the combined data sources of the FBS and the survey for each farm we investigated the current use of RT in England, the depth of tillage practices (given as average depth of the main tillage practice), where RT is used within the crop rotation and how tillage practices relate to key input and output data for individual crops.
The farms were aggregated into groups for analysis based on the following characteristics: farm size (based on the Defra's Standard Labour Requirement levels expressed in terms of full-time equivalents (FTE) was defined as: Small (<2 FTE), Medium (≥2 FTE and <3) or Large (≥3 FTE); farm type was classified as Cereals, General Cropping or Mixed arable and livestock, representative of the main arable farm types in England); location was identified both in terms of Government Office Region, GOR and European Union, EU, super region, which are, respectively, one of nine and one of three groupings within England); education level of the farmer; farmer age; and ploughing frequency before specific crops (i.e. 'ploughing always', 'ploughing sometimes' and 'ploughing never'). The percentage area of RT was determined using area weighting based on crop areas, GOR and farm type to aggregate results to be nationally representative of total area. The 'area weights' are derived from calculating the population areas of wheat, barley and oilseed rape, divided by the survey areas for crop, farm type and GOR. This provides area-weighting values by crop, GOR and farm type. These area weights are then applied to the individual farm results to provide nationally representative aggregate estimates (Glithero et al., 2013c).
Data were analysed using GenStat (VSN International Ltd.). RT use was compared among groups, after dividing farms based on whether they did or did not use RT practices, using chi-square with a 5% significance level. For all other analyses ANOVA was used at the 5% significance level after checking that the test assumptions were met. Variable costs, yields and gross margins were compared among groups varying in the frequency of ploughing for four crops. For comparison of tillage depth among groups, one farm, which had 100% RT, was excluded as it did not provide data on the depth of cultivation (it was possibly 100% direct drilled). The Bonferroni post hoc test (set at 10% significance level) was used after the ANOVA to determine significantly different groups.

Area of RT
Of the 249 farms surveyed, 113 (46%) used at least some form of RT; an additional 14 farms had used or were likely to use RT within the rotation, suggesting that over half of the farms surveyed used RT at some point. Based upon aggregation weighting, we identified that 32% of the arable land was subject to RT practices. For farms that used RT, the mean area under RT was 50%; however, the amount of RT on individual farms varied greatly (Figure 1). There were eight farms with >10% of the area under RT and 20 farms that did not plough at all.
Of the farm types, Cereal farms were more likely to use at least some RT (P < 0.001; data for the following analyses are presented in Table 2); however, where RT is used, farm types do not differ in the percentage area of RT (P = 0.548). Large farms were much more likely to use some RT than small farms (P < 0.001). For farms that used RT, farm size did not influence the proportion of land under RT (P = 0.090).
The number of farms using RT varied with GOR (P = 0.033) with the East Midlands and South East more likely to have some RT, whereas the North West, Yorkshire & The Humber and South West were more likely to have none. For farms where RT was used, the North East and Yorkshire & the Humber had significantly smaller areas of RT than the South East (P = 0.007). When these GOR regions were aggregated to EU super regions (East, West and North), farms in the East were more likely to have some RT than farms in the North and the West (P = 0.011). Where at least some RT is used, farms in the East had significantly greater areas of RT than those in the North (P = 0.001).

Depth of tillage
Average tillage depth of the main form of tillage for all surveyed farms was 20 cm. Shallower tillage depth tended to be associated with farms with all RT (P = 0.105; Figure 2) although their mean depth was only 3 cm shallower than the mean value when only ploughing was used. There was considerable variation in tillage depths, in particular for farms using all RT; for example, four respondents tilled to a depth of 0-10 cm, seven respondents tilled to a depth from 10 to 20 cm while eight respondents tilled to a depth over 20 cm. As there were only a small number of farms that had an average tillage depth of less than 10 cm it was not possible to identify specific characteristics relating to these farms.
Depth of tillage varied with farm type, being significantly deeper on General Cropping farms compared to Cereal and Mixed farm types (P < 0.001). There was a strong trend for deeper tillage on Large farms (P = 0.070) but the difference in average tillage depths between farm size groups was very small. Tillage depth was significantly greater in the East of England than in the South East (P < 0.001); however, when aggregated into EU super regions there was no significant difference (P = 0.193). There was no difference in tillage depth with level of education (P = 0.944) or age (P = 0.624).

Reduced tillage in the rotation
Analysing data only for farms that used some RT, the frequency of ploughing (i.e. whether they always plough [AP], sometimes plough [SP] or never plough [NP]) for different crops was considered both before the crop (Figure 3a) and after the crop (Figure 3b). Decisions about whether to plough before a crop will in part depend on the preceding crop; however, without knowledge of the specific rotations used on these farms it was not possible to explore this.
Reduced tillage is often used with winter wheat (WW) and winter oilseed rape (WOSR), less so with winter barley (WB), spring barley (SB) and winter field beans (WFB), and very little with root crops (i.e. potatoes and sugar beet; RC). As rotations are built around WW and WOSR, for farms with these crops it also tends to be used after these crops. For before and after WW, the majority of respondents recorded they sometimes plough; this was also a common response for WOSR and barley, demonstrating flexibility in the cultivation approach. Although WFB and RC are almost always planted after ploughing, WFB tended to be followed by RT while RC are almost always followed by ploughing.

Yields, gross margins and input costs
Yield and gross margin data for each crop were tested to assess the effect of frequency of ploughing (Tables 3-6). Farms were excluded that did not grow the specific crop or where gross margin data were not available; hence, the number of farm observation differs from those presented above. For farms that sometimes plough before a particular crop, we do not know if they ploughed or used RT; therefore, we take this as an indication that they adopt a 'flexible' management approach where tillage practice is varied to provide the most benefit. For WW, yield was lowest for AP farms (P = 0.089; Table 3). Spray costs were significantly greater for NP farms than AP farms (P = 0.006). GMs, fertilizer costs and seed costs do not vary significantly with the frequency of ploughing. There were no significant differences among the three ploughing frequency groups regarding input and output data for WOSR (Table 4). For WB, yields were significantly greater on SP farms (P = 0.008) while GMs were significantly lower for NP farms (P = 0.007; Table 5). Spray costs for NP farms were not significantly different from AP farms (P = 0.065). Seed costs were significantly greater for SP farms (P = 0.008). Fertilizer costs did not significantly differ with ploughing frequency. For SB yield, input costs and GM data did not vary significantly with frequency of ploughing (P = 0.062; Table 6).

Use of RT
Our estimate of 32% of land under RT is less than a previous contemporaneous UK estimate of 44% (Defra, 2010). Within Defra's (2010) Farm Practices Survey it is possible that a greater proportion of respondents with an   (Pennings et al., 2002). The current limited data available on RT practices throughout Europe appear to only capture the average aggregate RT use per country rather than data at the individual farm level. From our data, a small number of farms did not plough but the vast majority of farms using RT practices also ploughed (i.e. mixed tillage systems) largely rejecting the concept of a dichotomy of either using all RT or all ploughing. Typically mixed tillage occurs either at specific points within the rotation (rotational ploughing) or in response to specific conditions (strategic tillage), with rotational ploughing being the more frequently observed.
With respect to tillage practices and individual crops, for WW, the majority of farmers 'sometimes plough' prior to crop establishment; the major break crop, WOSR, which is less prone to weeds, can be established by broadcasting seed directly into the stubble from the previous crop (e.g. using an autocast system). RC and WFB were almost always ploughed beforehand. Van den Putte et al. (2010) found that RC yields were not significantly lower under RT; however, Lahmar (2010) notes that RCs are harder to manage under RT systems. WFB can be established via broadcasting and then incorporating seed directly with a plough, which is a quick method of establishment and suited to the wet soil conditions usually associated with later sowing. However, there is now growing interest in RT drilling of WFB (PGRO, 2015).
Our results indicate that tillage depth for RT, when it was the main tillage practice, was not significantly different from ploughing, contrasting with previous definitions of RT relating to less than 100 mm (Ingram, 2010) or 150 mm (Powlson et al., 2012). Hence, even in the absence of soil  inversion, substantial soil disturbance still occurs with RT. Cultivation depth may in part relate to the need to bury crop residues or weeds; indeed, Davies & Finney (2002) suggested that shallow RT was unlikely to become commonly used because it is associated with higher risk of yield penalties resulting from an increased grass weed burden. Van den Putte et al. (2010) also report that shallow RT depth can result in higher yield penalties.
Our results indicate that RT is most common on larger farms, supporting the findings of the SAPM. Large RT trains require large powered tractors (e.g. 310-400 horsepower; ABC, 2014), more easily financed on larger farms. Moreover, most of the farms recording RT use also undertook some ploughing; it is more feasible for larger farms to hold two types of cultivation equipment. RT may also be more suitable for larger farms due to their ability to cope with the risk burden associated with RT (Lahmar, 2010). It is likely that larger farms also have greater timeliness issues (Melander et al., 2013) giving them greater incentive to switch to faster crop establishment methods. Mixed farm types recorded lower use of RT, arguably resulting from the smaller non-RC areas, reducing the incentive to invest in larger RT machinery and tractors. A greater area of RT and a shallower depth of tillage was recorded in the South East, consistent with the greater number of Cereal farms relative to General Cropping farms and the observation of Cannell et al. (1978), which was based on broad soil characteristics, that large parts of the South East were suitable for ZT.

Tillage effects on yields, costs and the environment
The widespread use of RT identified from this study suggests that farmers are finding benefits from its use. The survey considered the impact of RT on variable costs and yields but based on the limited data, it is not possible to directly determine how beneficial RT practices are after considering fixed costs and dynamic effects. For example, while fuel costs are recorded at the farm level in the FBS, it is not possible to robustly compare fuel use between farms due to the variable use of contract services, which typically include the cost of fuel, versus own machinery use where fuel costs would be recorded separately. However, given these constraints, our results provide informative estimates at the GM level; specifically, GMs only significantly differed with frequency of ploughing for WB. Using RT may also allow other economic benefits; reducing the time required for field preparation could allow farmers to share tillage equipment, providing cost savings, or to undertake contracting work on other farms.
The literature suggests yields tend to be smaller under RT. However, our data do not support the view that for WW there is a yield penalty associated with the use of RT, although WB yields were smaller under RT. This latter result was not due to an interaction between RT usage rate and differences in crop yields among regions. Crop protection costs would be expected to be greater for RT systems because of the need for more of the weed control to depend on herbicides. WW had higher spray costs where RT was used; however, because crop protection cost data were not disaggregated into individual crop protection products, it was not possible to determine which sprays account for this increased cost. In Sweden, WW yields for RT tend to be much smaller when following another cereal crop (Arvidsson et al., 2014). Rieger et al. (2008) reported that in Switzerland the yield reduction resulted from diseases carried in crop residues, which would suggest that increased crop protection is required when WW follows another cereal. Interestingly, spray costs were not higher for 'never plough' before WB and SB; however, it is not possible to confirm whether this reflects a difference in behaviour among farmers (e.g. attitudes to the environment) or whether those that never plough have different rotations that reduce the need for crop protection products used on barley. Contrary to common assumptions, RT systems can be viable with low herbicide usage although this requires careful management, as seen with ZT experiments in South Portugal (Barros et al., 2007). Although spray costs were less for RT, seed costs tended to be greater, possibly reflecting a need for greater seed treatment or seed rate to combat weeds. Due to the modest sample size, it is not possible to attribute the extent of these cost differences to tillage practices. For the majority of the output metrics, there were no significant differences among regions; hence, it is unlikely that differences in RT practices were more frequently observed in regions with higher yields. Due to the constraints imposed by sample size, it was also not possible to separate farms based on the average depth of the main tillage type, so the effects of specific RT practices could not be identified.
Based on current RT practices, environmental benefits achieved with RT may not match those indicated in the literature. Morris et al. (2010) suggest that, depending on the equipment used, fuel requirements for some deep RT practices can be similar to those of ploughing, limiting greenhouse gas emission reductions. Without fuel-use data in our current study, it is not possible to assess this. Davies & Finney (2002) question whether the environmental benefits associated with RT seen in field experiments occur at the farm level; field experiments are unlikely to represent mixed tillage systems as observed in this study. Powlson et al. (2012) suggest increased soil carbon benefits from RT would be lost under rotational ploughing; however, it is unclear whether other environmental benefits from RT are achieved within the short term, within mixed tillage systems. It is likely that some RT tillage systems, such as ZT, are better suited to continuous use. In general, during the first several years after adoption of ZT crop yields are initially lower relative to ploughing but increase after about 3 yrs of use.
Although this could in part be due to time taken by the farmer to learn this new technique, one important factor is that it takes several years for the soil structure to improve suggesting that disruption to soil structure through rotational ploughing would prevent ZT from achieving yield parity with ploughing (Soane et al., 2012).
Farmers are encouraged to utilize tillage practices that minimize soil erosion; while the Soil Protection Review encourages RT, the EU's cross-compliance regulations encourage ploughing to avoid soil erosion caused by soil compaction (Defra, 2014). One aspect of RT that helps reduce soil erosion is maintenance of crop residue on the soil surface; however, the tillage depths recorded in this study suggest that little crop residue is left on the surface, which may arguably limit the soil erosion benefits of RT. Although the wider survey, from which these data were drawn, did collect information on straw use, due to the limited sample size it is not possible to compare tillage practices and straw use. As problems with residues in RT systems have been previously identified as less pronounced where straw is baled (Cannell, 1985), this might support the view that straw baling is more common on RT farms. However, one aim of RT practices is to increase soil organic matter, yet residue removal can negatively impact on soil organic matter levels (Blanco-Canqui & Lal, 2009). Moreover, contemporary residue management and RT techniques differ substantially from those of the 1980s. In Defra's (2010) Farm Practices Survey, RT is defined as having at least 30% crop residue cover but it is unclear whether this level of residue cover can be achieved with deep RT. 3 Considering the mean tillage depths recorded in our survey, it may be that the 44% of land under RT with 30% residue cover given in Defra (2010) is unrealistic.

Increasing the adoption of RT
There are arguments for encouraging greater use of RT practiceswider adoption would address some of the challenges raised under the banner of 'sustainable intensification' and the drive for greater productivity at lower environmental cost (Garnett et al., 2013) while meeting the demands of a growing, more affluent global population. Climate changeand hence greenhouse gas mitigationand soil erosion (Defra, 2009) are two areas where RT can play a part and both increasing the use of, and maximizing the environmental benefits from RT represent important interventions that can drive improvements in sustainable intensification. However, as we have seen, RT covers a wide range of practices and is suited to different types of farm situation (for example, the level of grass weed burden) and thus will not be adopted in all situations.
There are thus limits to the extent that RT can be applied in the UK; although Cannell et al. (1978) suggested approximately 80% of arable land could be used for ZT, Davies & Finney (2002) concluded that for successful RT, systems must be suited to specific soil types, site, scale and management factors, whereas ploughing is more universally applicable. Although our survey did not consider soil type and climate, the differences observed among regions and farm types support the argument that site-specific factors influence the applicability of RT. The corollary to this is that site-specific factors are likely to also limit the extent to which RT can be expanded.
Other factors may constrain further expansion of RT. The biggest concern regarding RT practices is reduced yields (Jones et al., 2006); farmers may also be reluctant to use RT due to the tradition of using the plough and concerns that, because RT practices tend to produce fields of less 'tidy' appearance, that their land will be perceived by neighbouring farms as being poorly managedthis is probably increasingly the case in the age of 'Google Earth'. There is also limited assistance for farmers wishing to switch to RT practices; The Soil Management Initiative Ltd (SMI), the UK branch of the European Conservation Agriculture Federation, was established to provide information on RT to farmers (e.g. SMI, 2005). But it is no longer available and there appears to be no organization providing this support.

Future research and industry recommendations
Our results present a current picture of tillage practice on commercial arable farms in England. Whereas previous research has focused on economic data for single farms, this study considered financial data across a number of farms to provide a better understanding of the variability among farms that utilize RT. While this did allow a better insight into financial impact of RT, it did necessitate a compromise in the comprehensiveness of data collection (e.g. no fuel-use data for individual crop enterprises). Future approaches to this type of research should explore the rotational and tillage practice effects in combination; for field-experimental observations, this will require longer-term studies to be undertaken.
Although providing information on RT use with different crops, the survey did not identify specific RT practices used or the adaptations farmers have made to incorporate RT into their farming systems. Tillage depth measurements were also limited to the average depth of the main tillage practice. Consideration of the specific characteristics of the farms, in particular soil type, would be required to better understand how farm-specific factors influence the outcomes of using RT. Future survey work should seek to determine the 3 In asking about arable land cultivation, the Farm Practices Survey did not provide an option for RT with less than 30% residue cover. reasons why farmers use RT in order to test the hypothesis that the benefits of RT as established in the literature correlate with commercial practice and observations. Further work is also needed to quantify short and longerterm environmental benefits to determine whether deep RT is environmentally beneficial. One potential barrier to future increases in RT is increased weed problems resulting from reduced effectiveness of available herbicides and stricter pesticide legislation limiting crop protection product availability (Melander et al., 2013). It is likely that broader changes to farm systems are required to maximize the potential of RT, for example by incorporating aspects of conservation agriculture, such as longer rotations.
Information on 'best practice' is limited and providing this information could assist farmers in adopting RT practices. Future research should explicitly generate applied, practical outcomes to facilitate further uptake of RT systems that maximize benefits to farmers and the environment.

Conclusions
There is widespread use of reduced tillage practices in England, although its use varied with farm size and type, region and specific crops grown. RT tended to be used only on part of the land demonstrating that many farmers practice a rotational ploughing system. The depth of soil disturbance suggests that the RT practices used in England are quite energy demanding, so the benefits of using RT are limited. Where RT is associated with genuine production and environmental benefits, policy makers seeking to achieve sustainable intensification objectives may choose to incentivize the use of RT practices. Our findings indicate that these incentives should address farm-level constraints such as the capital investment required to hold equipment for sustainable RT cultivation methods; recent (2015) changes in UK investment policy for farms give farmers an incentive to increase investment in new technology by raising the annual investment allowance but do not tie this to particular sustainable practices. Moreover, targeting increased uptake of RT in areas where RT use is already more prevalent, due to climatic or land-based factors, will facilitate policy success. Finally, in incentivizing RT uptake, policy makers need to be aware of the substantial soil disturbance frequently observed with on-farm RT use, which provides a potential limit to any environmental benefit that might be anticipated to result from increasing the use of RT in contemporary commercial practice.