1 Introduction

Grasslands cover on average 36.2% of the land area in the UK (Eurostat 2021), while grassland contribution to the total area farmed in Northern Ireland in 2019 was estimated as 79% compared to 4.4% of cropland (DAERA 2020). Agricultural grasslands across the UK remain important managed ecosystems that support livestock production and the delivery of multiple ecosystem services (ONS 2016; Qi et al. 2018). Managed grassland ecosystems receive significant amount of nutrient fertilisers to increase grass yields and many grasslands also experience regular soil tillage and reseeding events to rejuvenate the sward (Velthof et al. 2010), which could, however, negatively affect soils’ ability to store carbon (C) (Reinsch et al. 2018).

Current knowledge suggests that no-till permanent meadow systems promote soil C sequestration (Castelli et al. 2017) and that soil tillage can strongly influence soil organic carbon (SOC) dynamics and cause long-term SOC losses especially following increases in soil tillage frequency (Conant et al. 2007). There is, however, no overall consensus on whether tillage and reseeding of agricultural grasslands may, or may not, reduce soil C stocks in the long term depending, for instance, on the frequency of soil disturbance, ploughing and reseeding method, and soil type. Very few studies have specifically addressed how the frequency of ‘soil tillage + reseeding’ events (e.g. number of times that a grassland field was ploughed and reseeded in the previous 50 years) on soil C stocks (Fornara et al. 2020). Evidence from cropland systems, where no-till was compared to conventional tillage systems, shows, for example that no-tillage can promote SOC accumulation in topsoils (0–10-cm depth; West and Post 2002), or reduce amounts of SOC in deeper soils (Angers and Eriksen-Hamel 1997) or have no significant long-term effects on SOC stocks either in topsoil or subsoil (Dimassi et al. 2014). A recent comprehensive analysis of findings from 178 experimental sites worldwide suggests that there is high uncertainty (because of insufficient information) on the impact of no-till on SOC (Ogle et al. 2019).

Potential tillage effects on soil C dynamics in permanent grassland have been addressed after single ‘plough + reseeding’ events (Necpálová et al. 2014; Carolan and Fornara 2016; Drewer et al. 2017), and only assessed in the short term and rarely addressed in relation to the frequency of tillage events spanning many years of grassland management. Modelling predictions (Reinsch et al. 2018) suggest significant reductions of soil C stocks when grassland management includes frequent plough and reseeding events. Very little information exists, however, about how the frequency of tillage and reseeding events might affect soil C stocks in grasslands across ‘real-world’ farming systems, on different soil types and where multiple grassland management practices (e.g. nutrient fertilization, grazing) occur simultaneously.

The practice of tillage and reseeding is also carried out to improve soil bulk density because soil compaction can negatively affect grassland ability to deliver multiple ecosystem services (Newell-Price et al. 2013). Unfortunately, potential relationships between soil tillage frequency and long-term changes in soil bulk density and SOC stocks in grasslands remain poorly studied and understood. A previous study on croplands growing on deep-loess soils shows that bulk density was not affected by conversion of conventional tillage to no-tillage and did not change either between no-tillage and ridge-tillage in a continuous corn system (Logsdon and Karlen 2004). Another cropland study on heavy clay soils showed that bulk density was higher under no-tillage when compared to conventional tillage (Grant and Lafond 1993). Studies, which involved grasslands, did not find either significant differences in bulk density, for example along a chronosequence of reseeded grasslands (Breuer et al. 2006) or between grassland and cropland systems (Tan et al. 2004).

Variability in soil bulk density responses to tillage could affect soil C stocks and thus the assessment of soil C storage in agricultural grasslands. It could be that soil bulk density and SOC responses to tillage frequency in intensively managed grasslands might depend more on the use and frequency of heavy machinery, which can significantly increase soil compaction (Batey 2009) rather than on the frequency of tillage and reseeding per se.

The main goal of this study is to address whether and how soil C stocks might respond to repeated ‘soil till + reseeding’ events in intensively managed agricultural grasslands in Northern Ireland, UK, under wet and cool climate conditions. We first identified 500 grasslands fields across 46 farms using as main selection criterion the availability of information on management history of these grasslands especially in relation to reseeding frequency over the previous 50 years. Our expectation is to find a negative relationship between ‘soil till + reseeding’ frequency and soil C and N stocks of these managed grasslands.

2 Material and methods

2.1 Study sites description

To assess how the frequency of ‘soil tillage + grassland reseeding’ events might affect soil C stocks, we initially carried out a survey among >150 farmers/land owners in Northern Ireland between May and October 2019. The main aim of the survey was to search for detailed management history information on grassland fields especially in relation to the frequency of tillage and grassland reseeding events occurred in the previous 50 years (i.e. information going back to 1970). Based on the survey, we identified 46 farms and sampled 500 grassland fields for which information on the frequency of ‘tillage + reseeding’ events was available through existing records or based on best knowledge provided by the farmer. Grassland fields were identified across different soil types (Fig. 1) in Northern Ireland and additional information was gathered in relation to key management practices including (a) last (most recent) year of tillage + reseeding event, (b) land use management, (c) total nitrogen (N) added (inorganic + organic N), and (d) most recent liming application (Supplementary Table 1). These fields represent intensively managed grasslands, which all receive significant amount of nutrient fertilisers (organic and inorganic) and where ploughing and reseeding events are associated with the use of heavy machinery.

Fig. 1
figure 1

Distribution of 46 farms across Northern Ireland where 500 grassland fields were sampled between 2019 and 2020. Soil carbon sampling locations are spread across different soil types (World Reference Base (WRB) for soil classification).

Full size image

Grassland productivity has been measured across years through the GrassCheck research programme in Northern Ireland (https://www.agrisearch.org/grasscheck). Grass yields averaged for >20 years across farms throughout Northern Ireland including the soil types, management practices, and geographical locations of the 46 farms included in our study typically vary between 10.5 and 12.5 t DM ha−1 year−1. Note that most of this grass biomass is removed either through multiple grass cuts to produce silage and/or animal grazing.

The climate zone of Northern Ireland is characterised by mild-wet winters and by cool-humid summers with mean annual precipitation and temperatures across our lowland grassland sites varying between 800 and 1100 mm and 8.5 and 10 °C (https://www.metoffice.gov.uk/climate/uk/regional-climates/ni).

2.2 Soil sampling

In each of the 500 grassland fields identified across 46 farms, which were selected based on available information on ploughing and reseeding frequency, we collected soil samples at eight random locations within the 0–30-cm soil depth layer (following IPCC guidelines; IPCC 2006) using a 3-cm diameter soil corer. Soils collected were mixed to give one composite sample, transported to the laboratory, sieved through 2-mm mesh-size, air dried, ground, and analysed for total C% and N% concentrations using a LECO TruMac CN Analyser (St Joseph, MI, USA; https://knowledge.leco.com/analytical-library-cat/cns-analyzers/trumac-series). To determine inorganic C concentration, bulk soil subsamples were burned for 16 h at 550 °C in a muffle furnace and ashes analysed for inorganic C content, which revealed negligible inorganic C. Three undisturbed soil cores (5 cm in diameter) were also collected in each grassland field within the 0–30-cm soil depth layer to measure soil bulk density as the ratio between air-dried soil and soil volume.

2.3 Data analysis

Residual maximum likelihood (REML) estimation analysis was used to test for the potential effects of the frequency of reseeding on different response variables (e.g. soil C and N stocks, soil C and N (%) concentrations) with farm sites included as the random effect. Thus, REML was applied including reseeding frequency as fixed effect and blocking as Farm ID (i.e. 46 farms spread across Northern Ireland). Residual diagnostic was applied for detecting violation of normality assumption. The Wald test was performed to assess whether explanatory variables were significant in our model. Fisher’s LSD tests were performed to compare multiple means and the significance of the effect of different factors. We also performed simple and multiple regression analyses to identify potential relationships between management practices (i.e. land use, N additions, soil type, liming application, soil bulk density) on soil C content (%) and soil C storage. Analyses were conducted using VSN International (2020).

3 Results and discussion

3.1 Effects of tillage and grassland reseeding frequency on soil C and N stocks

Across the 500 grasslands sampled in this study, we found that tillage and grassland reseeding frequency did not affect soil C (F = 1.26, P = 0.28) or soil N stocks (F = 1.74, P = 0.17) measured between 0 and 30-cm depth (Fig. 2). Thus, contrary to our expectation, we found that soil C and N stocks within 0–30-cm depth were not significantly affected by the frequency of ‘tillage + reseeding’ events in the previous 50 years. These findings agree with those of a previous study carried across 126 grassland fields under the same Irish climate conditions (cool and humid), which did not show any significant effects of reseeding frequency on soil C and N stocks at different soil depths (0–20-, 20–40-, and 40–60-cm depth; Fornara et al. 2020). Evidence on how the frequency of soil tillage + reseeding events might affect soil C stocks in permanent grassland remains very limited. Findings from studies, which assessed long-term soil C stock responses to a single (one-time) till + reseeding event (Carolan and Fornara 2016) or a land use change (i.e. cropland to grassland; Breuer et al. 2006) across farmed landscapes, also suggest no significant effects of grassland ‘age’ on soil C stocks. Evidence from meta-analysis studies (reviewed by Dignac et al. 2017) across different agro-ecosystems also suggest that no-tillage may have little effects on SOC stocks.

Fig. 2
figure 2

Lack of a significant relationship between soil tillage and reseeding frequency and soil carbon (C) and nitrogen (N) stocks measured between 0- and 30-cm soil depth across 500 grassland fields.

Full size image

However, predictions, from modelling studies of the impact of the frequency of soil tillage on C stocks in grasslands (Reinsch et al. 2018) and croplands (Conant et al. 2007), both suggest that soil C storage will be reduced by increasing tillage frequency. Findings from our study (see Fig. 2 and Fig. 3) suggest that when the frequency of ‘tillage + reseeding’ increases, soil C stocks and soil C (%) content become less variable and tend to be negative (e.g. soil C storage potential tends to decrease). Our study also suggests that long-term changes in soil C stocks could be better predicted if fluctuations in soil bulk density are incorporated into simulation models, which aim to assess potential interactions between soil structure and the storage and turnover of soil organic matter (Meurer et al. 2020).

Fig. 3
figure 3

Negative significant relationship between soil tillage and reseeding frequency and soil carbon (C) content (%) and nitrogen (N) content (%) measured between 0- and 30-cm soil depth across 500 grassland fields. Best fitted lines: Soil C (%) = 6.71 − 0.226x; Soil N (%) = 0.55 − 0.01×.

Full size image

In general, there is still high uncertainty associated with potential long-term impact of tillage on C stocks (Ogle et al. 2019), whereas there is more consensus on the short-term impact of tillage, which tends to significantly increase soil C losses across different agroecosystems (Reicosky et al. 1997; Necpálová et al. 2014; Abdalla et al. 2016; Carolan and Fornara 2016; Reinsch et al. 2018). Contrasting long-term soil C stock responses to the frequency of tillage + reseeding could be explained by differences in other factors such as management intensity, soil type, ploughing method, nutrient fertilization regime, and climatic conditions. In our study, we did not find any significant effect of soil type or the total amount of N applied (inorganic + organic N) on soil C% or soil C stocks (P > 0.5). It could be that specific grassland management effects on soil C storage have been masked by the interaction of different practices that have been occurring simultaneously across many years. Nevertheless, it is important to investigate what biogeochemical mechanisms or processes might explain the lack of a clear relationship between soil C storage and the frequency of tillage and reseeding in permanent grassland as we found in our study.

3.2 Effects of soil bulk density, N fertilization, and soil till and reseeding on soil C (%)

We addressed what factors might explain no changes in soil C stocks across these agricultural grasslands and found that soil C content (%) was negatively affected by soil till and reseeding frequency (F = 13.3, P < 0.0001; Fig. 3). Similarly, soil till and reseeding frequency negatively affected soil N content (%) (F = 22.6, P < 0.0001; Fig. 3). Thus, these results would suggest that soil C% and N% tend to decrease at increasing frequency of soil disturbance events, and we performed further statistical analyses to search for potential links between changes in soil C% (and N%) and other soil variables and grassland management factors. First, we found a positive significant relationship between tillage and reseeding frequency and soil bulk density (F = 39.4, P < 0.0001) also shown by the results of the regression analysis in Fig. 4 (R2 = 0.12, P < 0.0001). Second, when we performed multiple regression analyses where different management practices were simultaneously included, we found that the most important variable affecting soil C (%) was soil bulk density (Table 1) whereby increases in bulk density (i.e. increases in soil compaction) were associated with lower soil C content (Fig. 5). Similarly, we found that soil N (%) was significantly affected by soil bulk density but not by N fertilization rates as estimated across the 500 grassland sites. Linear regression analyses did not show any significant effects of nutrient fertilization (i.e. total organic or inorganic N fertilization) on soil C (%), soil C stocks, soil N (%), or soil N stocks. This could be because these are intensively managed grasslands, which receive significant amounts of N fertiliser both in inorganic and organic forms and could be that soil N availability is not limiting (N is actually in excess) for biomass production and the formation of soil organic matter which requires N and C binding together.

Fig. 4
figure 4

Positive significant relationship between soil tillage and reseeding frequency and soil bulk density measured between 0- and 30-cm soil depth across 500 grassland fields. Best fitted curve: Soil bulk density = 0.76 + 0.027× − 0.001*(× − 2.75)2.

Full size image
Table 1 Dependence of soil C (%) on soil bulk density (g cm−3), total N fertilization (including organic and inorganic fertilization; Mg N ha−1 year−1), and soil till and reseeding frequency. Results refer to a multiple regression analysis (overall R2 = 0.37).
Full size table
Fig. 5
figure 5

Negative significant relationship between soil tillage and reseeding frequency and soil carbon (C) content (%) measured between 0- and 30-cm soil depth across 500 grassland fields. Best fitted curve: Ln (Soil C) = 1.498 − 1.09*Ln (Soil Bulk density).

Full size image

Our findings show how soil C (%) and N (%) concentrations significantly decreased while soil bulk density increased with the frequency of soil till + grassland reseeding events. Because soil C stocks are calculated by multiplying C (%) with bulk density, the opposite response of these two key factors has contributed to ‘homogenise’ the soil C stock product possibly partly explaining the lack of a significant relationship between soil C stocks and the frequency of soil till + reseeding events. In this case, younger (more frequently reseeded) grasslands would have lower C (%) but higher bulk density than older (less frequently reseeded or permanent) grasslands. These contrasting trends in soil C (%) and bulk density explain the observed ‘homogenization’ of soil C stocks across the range of grasslands included in our study. These results, and in particular the relationship between soil C (%) and bulk density, also provide some explanation on the relative role of the frequency of soil till + reseeding and other factors in affecting soil C and N stocks in intensively managed grasslands.

Our results suggest that any increase in soil compaction my reduce the soil C storage potential of intensively managed grasslands and possibly impact soil ecosystem functioning as also suggested in a previous study of UK agricultural grasslands (Newell-Price et al. 2013). Thus, soil C storage potential of grassland fields, which are reseeded at the same frequency, will be negatively affected more in fields associated with greater soil compaction. These results agree with those from a 60-year chronosequence study focused on the conversion of cropland to grassland across farmed landscapes in Germany, which suggests that bulk density is the most important soil property to predict changes in soil C and N concentrations (Breuer et al. 2006). A similar message was suggested by Carolan and Fornara (2016), which showed that increases in bulk density were more important than grassland ‘age’ in affecting soil C stocks possibly because greater bulk density was related to reduced C pools of small soil physical aggregate fractions. Our evidence is that bulk density increases with the frequency of tillage and reseeding events, which could be related to the more frequent use of heavy farm machinery. A study of 300 managed grasslands across England and Wales (Newell-Price et al. 2013) shows that increases in the number of machinery days had a significant negative impact on soil structure suggesting a positive relationship between the frequency of trafficking and bulk density.

3.3 Practical implications for grassland management

Our study focused on intensively managed agricultural grasslands under cool and humid climate conditions bringing evidence that increases in the frequency of tillage and reseeding events may not affect soil C or N stocks between 0- and 30-cm depth in the long term. An important finding of our study is that decreases in soil C (%) concentrations with increasing frequency of tillage and reseeding may not be caused by the frequency of reseeding per se but by increases in bulk density.

A first practical implication of our study is that the use of heavy machinery (e.g. increased number of machinery days) during reseeding and subsequent forage harvesting operations may be detrimental to soil C (%) because of its negative effects on bulk density. Our findings support those of Newell-Price et al. (2012) across 300 grasslands in England and Wales, which show a general pattern of lower soil compaction (lower bulk density) in older swards (permanent grassland) compared to younger grasslands and that bulk density generally increased with the number of machinery passes, at least on mineral soils as we found in our study. Thus, our findings suggest that increased awareness on the impact of trafficking during reseeding practices could help reducing the frequency of machinery passes and in turn improve bulk density. Going forward the introduction of higher precision and lightweight machines controlled by information and communication technology may help relieving mechanical pressures on soils (Techen and Helming 2017).

Another practical implication is associated with the potential impacts of managing grasslands mainly for grazing animals or silage cutting. We compared grassland fields, which were mainly grazed with fields which were used for silage, and we found that land management has significantly affected soil bulk density and soil C (%) content and soil C stocks. Specifically, we found that soil bulk density was significantly lower (F = 9.67, P = 0.002) in grasslands which are mainly grazed than grasslands managed for silage (Fig. 6). On the contrary, soil C stocks (F = 4.7, P = 0.03) and soil C (%) content (F = 10.2, P = 0.001; Fig. 6) were significantly higher in grasslands which are mainly grazed than grasslands managed for silage. Results from a multinomial logistic regression show how the predicted probabilities of greater soil till and reseeding frequency decrease when grasslands are managed mainly for grazing rather than for silage cut (Fig. 7). For example, the probability of finding a grazed grassland that is tilled and reseeded every 5 years is 10% compared to 60% for a grassland managed for silage. Thus, evidence from this study suggests that predicted probabilities of greater soil till and reseeding frequency decrease when grasslands are mainly managed for grazing rather than for silage cut. Most grasslands included in our study are grazed only for certain periods during the year and may not be referred to as under ‘continuous grazing’. Nevertheless, our evidence support findings from previous studies, which suggest that suitable management of grazing animals can contribute to increases in soil organic C levels (e.g. Soussana and Lemaire 2014; Conant et al. 2017). Grazed grasslands may benefit from greater organic C and N inputs to soils from animal wastes, and the recycling and processing of this organic matter by the soil microbial community can contribute to the accumulation of organic C more than in grasslands only used for silage production. Further research is needed to better understand how managed grasslands can actively sequester C under different soil types and climate conditions with the aim to develop suitable management strategies, which could promote the delivery of multiple ecosystem services (EIP Agri Focus Group Grazing for Carbon 2018). For example, non-frequent pasture renewal events involving full inversion tillage (FIT) may contribute to increase soil C accumulation potential in topsoils by bringing low-C subsoils to the surface (Lawrence-Smith et al. 2021).

Fig. 6
figure 6

Response of soil bulk density (g cm−3) (top) and soil C (%) (bottom) to land use management for grasslands which have been mainly a grazed, b cut for silage, or c both in the previous 50 years. Different letters indicate significant differences in response variables to land management.

Full size image
Fig. 7
figure 7

Predicted probabilities of tillage and reseeding frequencies of each land use management (grazing, silage and silage/grazing) based on data for 500 grasslands identified across Northern Ireland. Legend: Crosses—Silage/Grazing; Open circles—Silage; Closed circles—Grazing.

Full size image

4 Conclusion

Overall, our study brings evidence that increases in the frequency of tillage and reseeding events may negatively affect soil C stocks in the long term when soil compaction (i.e. bulk density) also increases, possibly reducing the soils’ potential to accumulate more organic C. Thus, increases in bulk density may be more important in affecting soil C stocks than the frequency of tillage and grassland reseeding per se. We would not suggest, however, to reseed these grassland soils more frequently than every 10 years because more frequent tilling and reseeding may not only have negative effects on soil bulk density but may also contribute to higher C losses from greater soil CO2 fluxes in the short term and possibly to greater nutrient losses from soils. Our findings show that when reseeding frequency increases (i.e. grasslands tilled and reseeded every 8, 7, or 5 years), reductions in soil C storage potential become more consistent and less variable. Finally, our study suggests that managing grasslands as pastures and promoting grazing such as ‘rotational grazing’ (or managing silage production to reduce the number of machinery passes) may contribute to increase soil organic C accumulation potential, which can benefit from additional C and N inputs through animal wastes and/or reduced soil compaction.