Environmental performance of crop cultivation at different sites and nitrogen rates in Sweden
Nitrogen (N) fertilisation has positive and negative effects on the environmental impact of crop cultivation. The mechanisms governing these effects are highly site-dependent, a factor often ignored in assessments of the environmental impact of crop cultivation. By assessing outputs of crop rotations using a life cycle approach, this study explored how greenhouse gas emissions and marine eutrophication caused by crop cultivation (including upstream processes such as production of farm inputs) depend on fertiliser rate and the site at which the cultivation occurs. Cereal unit (CU) was used as the functional unit. The calculations were based on data from multi-site long-term field experiments in Sweden and site-dependent data and models for non-measured processes. Cultivation at three N levels was evaluated, where the highest N rate was close to current average practices and the lowest level corresponded to one-third of that. Site characteristics had a stronger influence on both greenhouse gas emissions and marine eutrophication (variations of up to 330% and 490%, respectively, within N levels) than N level (variations of up to 74% and 59%, respectively, within sites). Main sources of variation in greenhouse gas emissions were soil nitrous oxide emissions (58–810 g CO2eq CU−1) and soil organic carbon changes (14–720 g CO2eq CU−1), while variations in marine eutrophication were mainly explained by field-level waterborne N losses (0.9–8.2 g Neq CU−1). The large variation between sites highlights the importance of considering site characteristics when assessing the environmental impact of crop cultivation and evaluating the environmental consequences of crop management practices.
KeywordsEnvironmental impact Life cycle assessment Eutrophication Climate impact Greenhouse gas emissions Site-dependent
The introduction of mineral nitrogen (N) fertilisers has enabled yield increases and has been a crucial factor for increasing global food security during the past century (Sutton et al. 2013). Increasing crop yields may also have several positive effects on the environment, such as increasing resource use efficiency, avoiding land use change and increasing input of crop residues available for soil carbon sequestration, but may require increased inputs of e.g. water and pesticides (Snyder et al. 2009; Tilman et al. 2002). Furthermore, N fertiliser use itself causes negative environmental impacts through fossil fuel use and losses of reactive N both during the production phase and after field application. It is therefore possible that there are trade-offs between several environmental objectives. Due to its multifaceted effects on agroecosystems, N fertiliser rate in relation to yield often emerges as one of the most influential variables for the magnitude of environmental impact of cropping systems (Van Stappen et al. 2017; Aoun et al. 2014; Charles et al. 2006). However, the relationship between N fertiliser rate and environmental impacts is complex, since both yield and field emissions of reactive N exhibit non-linear responses to increased N fertiliser rate (Delin and Stenberg 2014). Since both yield and emissions are also affected by local cultivation conditions such as climate and soil characteristics (Delin et al. 2005; Rochette et al. 2018), it is possible that the environmental impact of cultivation under different fertiliser rates will depend on the site. In that context, multi-site long-term field experiments are a valuable source of data.
Nitrogen fertiliser use is closely related to both greenhouse gas (GHG) emissions and nutrient losses to watersheds, which have been identified as key areas for actions to decrease the environmental impact of agriculture in the EU (European Commission 2017). Previous studies have shown that decreasing fertiliser input can reduce GHG emissions per unit produced (e.g. Ashworth et al. 2015; Williams et al. 2010; Brentrup et al. 2004; Goglio et al. 2012). However, since N fertilisers promote higher yields, this outcome may change if soil organic carbon (SOC) changes are included, which are also affected by site conditions (García-Ruiz et al. 2019). Another potential source of spatial variability in estimated GHG emissions is soil nitrous oxide (N2O) emissions, which are highly dependent on site conditions (Goglio et al. 2018). Site can thus influence estimated GHG emissions even though the impact itself is site-independent. Unlike climate impact, eutrophication is a site-dependent impact that mainly occurs due to regional emissions (Potting and Hauschild 2006). Eutrophication is a particularly urgent environmental issue for Sweden, which is surrounded by the world’s largest anthropogenic hypoxic zone, the Baltic Sea (Diaz and Rosenberg 2008), where agriculture is one of the main contributors to this issue (Licker et al. 2010). The site-dependent nature of the eutrophication cause-effect chain means that it is necessary to consider where emissions occur in order to make meaningful assessments of eutrophication impact, especially when comparing products produced at different sites (Page et al. 2014).
A system perspective is necessary to account for the environmental impacts of different substance flows occurring during the process of producing crops. A commonly used tool for determining the environmental impact of a product is life cycle assessment (LCA), which is frequently used for agricultural products (Notarnicola et al. 2017). In LCA, emissions and resource use during the whole or part of the life cycle of a product or process are compiled and related to environmental impacts (ISO 2006). This approach enables identification of possible trade-offs between products or production process alternatives, and thereby reveals potential burden shifts between impact categories or life-cycle stages. LCA of cropping systems is typically performed for either a site representing average conditions or for one or a few specific sites, due to lack of data. Different methodologies and degrees of spatial specificity limit inter-comparability (Bessou et al. 2013), and it is often unclear whether differences in results between studies can be attributed to spatial or methodological variations. The aim of this study was therefore to evaluate cradle-to-farm gate GHG emissions and marine eutrophication of crop cultivation at different N fertiliser levels and different field sites in Sweden, and to use the results to analyse the influence of these two factors on estimated environmental impact. Data from multi-site long-term field trials were used, together with site-dependent data and models for non-measured processes.
Experimental sites and set-up
Crop rotations and fertiliser rates in the southern and central cropping systems
Spring oilseed rapeb
Spring oilseed rapeb
Average fertiliser rate at low N level
Average fertiliser rate at medium N level
Average fertiliser rate at high N level
Goal, scope, functional unit, system boundaries and assumptions
Inputs (fertiliser and pesticide production)
Field operations (fuel consumption, production and maintenance of machinery)
Soil emissions (waterborne N and P compounds and airborne N2O, ammonia (NH3) and NOx)
In LCA, a functional unit is needed to relate inputs and outputs to a common unit. The functional unit is a quantified description of the product system functions (ISO 2006), and should be chosen with respect to the goals of the study (Caffrey and Veal 2013). Dry or fresh matter mass of harvested crop, together with area occupied by the crop, are the most common functional units used for agricultural LCAs (Notarnicola et al. 2017). However, mass is not a fair representation of the crop function when different crop types are present, and therefore cereal unit (CU) was chosen as the functional unit in this study instead. Cereal unit was first introduced to LCAs by Brankatschk and Finkbeiner (2014), as a basis for co-product allocation between grain and straw (ibid.) or between the crops in a crop rotation (Brankatschk and Finkbeiner 2015). It has also been used as a functional unit in LCAs comparing cropping systems (Prechsl et al. 2017). The CU represents the animal feeding value of each crop related to a reference crop (barley), and thereby enables crop yields to be represented by a unit that also considers crop function. One CU corresponds to 1 kg of fresh matter barley (at 14% moisture content). The animal feeding value is calculated based on protein, lipid, fibre and carbohydrate content of the crop, combined with data on animal species composition and species-specific metabolism. We used the CU values presented in the Supplementary Material of Brankatschk and Finkbeiner (2014), where calculation procedures are also presented in more detail. The relationship between harvested dry mass and CU harvested from the trial plots during the study period (1988–2009) is presented in Fig. 2. Subdivision of inputs and emissions among the crops in the rotation was not necessary in this study, since the aim was to assess the environmental impact over the whole crop rotation. Inventories were instead calculated for each treatment and year in the study period, and then aggregated over all years in the study period and divided by the amount of CU produced for each treatment during the study period. Amount of land occupied by crop cultivation was also determined by calculating the field area used per CU in each treatment.
Activity and emissions data for life cycle inventory
Most of the farm activity data were obtained from field measurements at the trial sites. They included data on harvested yield, soil carbon content, bulk density and soil texture. The study period chosen was 1988–2010, since data were readily accessible and consistent crop rotations were used for all sites during this period. Yields from 4 years in each time series were missing due to harvest failure, and these years were thus removed from the series at all sites in the two regions, to ensure consistent number of occurrences of each crop in each of the two crop rotations (see Table SM2 in Supplementary Material). Data that were not available from the field experiments, such as field operations, fertiliser types and pesticide amounts, were chosen to represent typical Swedish farming practices and, as far as possible, region-specific data were used. Specific data for all activities included are presented in this section.
Names of ecoinvent processes used for inventory data in this study. All inventories were taken from ecoinvent version 3.3, allocation: cut-off by classification
Ecoinvent process name
Used in this study for:
Pesticide production, unspecified, RER
Pesticide production emissions
Tillage, cultivating, chiselling, RoW
Machinery required for stubble cultivation
Tillage, ploughing, RoW
Machinery required for ploughing
Tillage, harrowing, by spring tine harrow, RoW
Machinery required for harrowing
Machinery required for sowing
Fertilising, by broadcaster, RoW
Machinery required for applying fertiliser
Application of plant protection product, by field sprayer, RoW
Machinery required for applying pesticides
Combine harvesting, RoW
Machinery required for threshing
Harvesting, by complete harvester, beets, RoW
Machinery required for harvest/baling of sugar beet
Agricultural machinery production, unspecified, RoW
Production emissions per unit machinery
Emissions from machinery production were calculated from the number of passes (see Table SM6 in Supplementary Material), data on kg machinery consumption per pass for each operation and emissions data for production of agricultural machinery from the ecoinvent database (Table 2).
Waterborne N emissions at field level (nitrate leaching) were calculated using a model implemented in the farm management tool VERA, developed by the Swedish Board of Agriculture, adapted to Swedish conditions and mainly based on empirical data. The model is described by Aronsson and Torstensson (2004) and calculates N leaching using site- and management-dependent typical leaching, and with correction factors for over- or under-fertilisation in relation to the yield achieved. Further information on the model structure and parameters used is provided in Supplementary Material to this paper. Phosphorus losses at field level were assumed to be equal to the region-, soil- and crop-specific average P losses, using data from Johnsson et al. (2016).
The cereals in the rotations were assumed to be harvested at 17% moisture content at sites S1–S4 and 19% moisture content at sites C1–C5, and dried to 14% moisture content. Oilseed rape was assumed to be dried from 12 to 9% moisture content, while no drying or other post-harvest treatment of sugar beet was included, since this typically occurs after farm-gate. The energy consumption during crop drying was assumed to be 0.15 L of heating oil per kg of water removed and 19 kWh of electricity per ton grain (Edström et al. 2005). Life cycle emissions data for heating oil were taken from (Gode et al. 2011) and data on the Swedish electricity mix from (Brander et al. 2011).
Organic carbon concentration in the topsoil (0–20 cm) was measured 6–10 times in every treatment during the whole trial lifetime (1956–2011), and we opted to use these data to derive mean SOC changes. This was done by estimating the slope for SOC concentration measurements over time for each treatment using linear regression analysis (see regression plots in Fig. SM1). Annual carbon stock changes in the topsoil were then estimated from the slope and site-specific bulk density values. The choice to include a longer time period for the SOC change estimation was made to avoid unfair allocation between years in the study period and because there were too few measurements during the study period. This should be considered when interpreting calculated SOC changes, and for this reason we report and discuss both total estimated GHG emissions and GHG emissions from non-SOC processes.
Seed production was included by subtracting corresponding seed rates from yield. Assumed seed rate for barley was 170 kg ha−1, for wheat and oats 210 kg ha−1 and for oilseed rape 6 kg ha−1 (Ahlgren et al. 2009). It was assumed that 60 kg sugar beet were required to produce the 2.1 kg seeds needed per ha (Nemecek and Schnetzer 2011a, b).
Liming rates in the experiments were very low (on average 0.1–0.15 kg CaO ha−1 year−1) during the study period and it was therefore not considered relevant to include impacts of liming.
Life cycle impact assessment
The GHGs carbon dioxide (CO2), methane (CH4) and N2O were included in the assessment, and their climate impacts were calculated using the IPCC characterisation factors (including climate-carbon feedbacks) for GWP100 reported by Myhre et al. (2013).
The marine eutrophication impact of waterborne N and P losses at field level was calculated using the method described by Henryson et al. (2018). This method uses site-dependent characterisation factors, which express the impact in N equivalents (Neq) and account for retention of nutrients along their way to marine environments, as well as the limiting nutrient in the marine recipient, while most other marine eutrophication life cycle impact assessment (LCIA) methods only account for N additions. In contrast to the majority of marine recipients, the sub-basins of the Baltic Sea surrounding Sweden can be growth limited by either N or P, or both (Swedish EPA 2006). Impacts of other N emissions were calculated using the ReCiPe2008 LCIA method (Struijs et al. 2013). Waterborne P emissions during fertiliser production were assumed to occur at Yara’s Siilinjärvi phosphate mine in Eastern Finland, which is Europe’s only phosphate mine and thereby the only source from which P-containing production effluent could end up in the Baltic Sea. This assumption therefore represents the worst case in terms of the marine eutrophication impact of P emissions during fertiliser production. Retention of 98% was assumed for P-containing effluent from Siilinjärvi (Huttunen et al. 2013). All other P emissions were assumed to be emitted at locations where they would not reach a P-limited marine environment and were therefore not included.
In addition to the ANOVA, a simple sensitivity analysis was performed for direct soil N2O emissions and waterborne soil N emissions, since these were modelled and expected to contribute substantially to the estimated impacts. The sensitivity analysis was performed by changing the modelled direct soil N2O emissions and waterborne N emissions one at a time by 20%.
Statistical analyses was performed using the software R version 3.5.1. All results from statistical tests are provided in the Supplementary Material (Tables SM9–SM12).
Mean total GHG emissions were 620, 560 and 600 g CO2eq CU−1 for the low, medium and high N level, respectively, or 380, 400 and 480 g CO2eq CU−1, respectively, when SOC changes were excluded (Fig. 3). However, differences between N level means were not statistically significant for total GHG emissions, and differences between low and medium N level means were not statistically significant when SOC changes were excluded (Tables SM10–SM11). Mean GHG emissions for the sites varied between 290 and 1090 g CO2eq CU−1 for total GHG emissions, or 220 and 800 g CO2eq CU−1 when SOC changes were excluded (Fig. 3). In contrast to the differences between N level means, the differences between site means were highly statistically significant for both total GHG emissions and non-SOC GHG emissions (Table SM10).
Marine eutrophication varied from 2.0 to 9.9 g Neq CU−1 (Fig. 5), with a maximum difference of 300%, 330% and 390% between sites at the low, medium and high N levels, respectively, and a maximum difference of 7–59% between N levels at different sites. Thus, similarly to the results for GHG emissions, differences between N levels were generally smaller than differences between sites. The medium N level gave the lowest marine eutrophication at all central sites (C1–C5) and one of the southern sites (S3), while the low N level gave the lowest marine eutrophication at the remaining three southern sites. The high N level gave the highest impact at five of the sites and the lowest N level at the remaining four sites. Mean marine eutrophication was 4.4, 4.2 and 5.0 g Neq CU−1 for the low, medium and high N level, respectively (Fig. 5), but only the difference between medium and high N levels was statistically significant (Table SM11 in Supplementary Material). In contrast, differences between site means were highly statistically significant (Table SM10) and varied between 2.1 and 9.1 g Neq CU−1.
The sensitivity analysis showed that altering soil N2O emissions by 20% changed estimated non-SOC GHG emissions by 3.0–15% (average 8.3%) for the respective treatments. The effect on total GHG emissions was slightly smaller (range 1.1–12%, average 6.3%). Changing waterborne field N emissions by 20% had a smaller effect on both non-SOC GHG emissions (range 0.2–1.7%, average 1.3%) and total GHG emissions (range 0.3–3.3%, average 0.8%). However, it had a larger effect on estimated marine eutrophication, which changed by 6.7–19.6% (average 16%) for the respective treatments. Full results of the sensitivity analysis are available in Table SM12.
Land area occupied by crops was inversely proportional to the amount of CU produced and therefore declined with higher N level. However, the rate of decline was smaller between medium and high N levels than between low and medium N levels (Fig. 3). Mean land occupation was 3.0, 2.3 and 2.1 m2 year CU−1 for low, medium and high N level, respectively. Mean land occupation for the sites varied between 1.6 and 3.9 m2 year CU−1.
Greenhouse gas emissions
The estimated GHG emissions differed between sites, both in terms of differences between N levels and in terms of magnitude. A feature in common for the sites which had the lowest non-SOC GHG emissions at medium N level (S3, C1 and C2) instead of low N level was that they showed a relatively high difference in yield between the low and medium N level (Fig. 2). This yield difference was thereby sufficiently large to outweigh the increase in emissions per unit area. Among the southern sites, high-yielding sites (S1 and S3) had lower GHG emissions than low-yielding sites, whereas this was not observed for the central sites, where the low-yielding site C3 had the second lowest non-SOC GHG emissions (Figs. 2 and 6). One of the reasons for this was the magnitude of yield differences between low- and high-yielding sites (Fig. 2). Another reason was that the high-yielding central sites (C1 and C2) had relatively high soil N2O emissions per hectare, which in turn was mainly due to their low sand content. There was no consistent difference in the magnitude of GHG emissions between southern and central sites, which indicates that soil characteristics may be more important for GHG emissions than geographical location, at least within the ranges of N fertiliser rates considered here.
The major contributing factor to the variation in non-SOC GHGs was soil N2O emissions. Field-level N2O emissions have previously been shown to be the main contributor to GHG emissions from crop cultivation (Goglio et al. 2014), and this was confirmed by our results to a certain extent, although the importance of soil N2O emissions differed between treatments (26–78% of total non-SOC GHGs; Fig. 3). Differences in soil N2O emissions were also larger between sites (up to 900%) than between N levels (up to 90%). However, the most common approach for calculating soil N2O emissions in LCA is to use default emission factors from IPCC Tier I methodology (De Klein et al. 2006), which neglect all local factors such as soil characteristics and weather and instead assume that 1% of the applied mineral fertiliser N is emitted as direct N2O, regardless of other factors. Since site characteristics have been shown to be highly important for N2O emissions (Rochette et al. 2018), the IPCC Tier I emission factors have been shown to be a poor predictor of N2O emissions at field scale (Goglio et al. 2018). Instead, we used a model derived from meta-analysis of field measurements in Canada, where for example precipitation was identified as an even more important variable than N rate supplied (Rochette et al. 2018). Despite not being verified on a global scale, this model should give better estimates than global emissions factors of field-level N2O emissions in cold temperate regions like Sweden. Applying the field data-based model to our treatments gave emissions factors of 0.4–6.3% (fraction of added N fertiliser emitted as direct N2O), with variations between sites and N levels. The results highlight the importance of using site-dependent emission models when comparing spatially dependent processes at different sites, as recommended by others (Peter et al. 2016) but still rarely used for estimating soil N2O emissions in LCAs.
Correlations and trade-offs between greenhouse gas emissions and marine eutrophication
Mean impacts were lowest at the medium N level for both GHG emissions and marine eutrophication (Fig. 6), but only two individual sites (S3 and C1) had their lowest impact at the medium N level for both GHGs and marine eutrophication (Figs. 3 and 5). This outcome highlights the importance of choosing an appropriate spatial scale with respect to the question to be answered in order to attain relevant LCA results. Similarly, previous studies have found that impacts were smallest at the lowest (Ashworth et al. 2015; Goglio et al. 2012; Brentrup et al. 2004), medium (Ruan et al. 2016; Wang et al. 2016) or highest (Charles et al. 2006) N fertiliser rate tested in each study. While the methods, models and assumptions differ between these studies, site-related differences may also explain why they reach different conclusions.
Despite climate impact gaining most of the attention as the dominant environmental issues of our time, waterborne reactive N losses are estimated to impose greater costs in the EU, through damage to human health and ecosystems, than the climate impact caused by N2O emissions (Sutton et al. 2011). As Fig. 6 illustrates, there were trade-offs between GHG emissions and marine eutrophication at some sites, which means that aspects that decreases GHG emissions does not necessarily decrease marine eutrophication. Although the major process contributions to both impact categories (soil N2O emissions, SOC changes, waterborne N and P losses) are dependent on the same factors (soil characteristics, climate, fertilisation, yield), they were clearly influenced in different ways at different sites. For example, sand content is reported to be a reducing factor for N2O emissions, while a high soil sand content increases the risk of N leaching (Rochette et al. 2018; Kyllmar et al. 2006). This effect is evident for the sandy soil at site C3, where soil N2O emissions were low while N leaching was high (Figs. 3 and 5). However, the connection to soil texture was not consistent for all sites, since both N2O emissions and N leaching per unit produced are also affected by other factors such as climate and yield response. For example, the increase in both GHG emissions and marine eutrophication from medium to high N level at site S2 and site S4 can be explained by the weak yield response (Fig. 2), which in turn may be an indication that factors other than N fertiliser limited growth in those treatment.
Data from long-term field trials are a valuable resource for life cycle assessments, since the availability of empirical data decreases the dependence on models and their associated uncertainties. However, using measured data also means that variations due to e.g. crop disease and extreme weather events are embodied in the results. There are also uncertainties connected to measuring methods, which could be one reason for the large uncertainties in estimated SOC change (Fig. 4). A similar pattern where field data exhibited larger variability than modelled data was observed by Goglio et al. (2018). Although the farming practices used in the Swedish long-term field trials represent typical farming practices in terms of crop rotation, field operations and amount of inputs, a typical farmer would adapt e.g. crop choice and annual fertiliser amounts to yearly conditions and cumulative experience, instead of leaving cropping system management unchanged over more than 60 years. For that reason, the quantitative results from this study should not be interpreted as a benchmark for average Swedish crop cultivation, or as representing the environmental impact of an ideal cropping system at each individual site. On the other hand, consistent management at the different sites over time enabled differences arising from site-dependent characteristics to be distinguished.
Despite LCA being an ISO standardised methodology (ISO 2006), performing an LCA requires making methodological choices, which could significantly affect the outcomes of the study. One such choice is the functional unit, which is particularly complicated for agricultural systems due to their multiple functions and outputs (Brankatschk and Finkbeiner 2015; Notarnicola et al. 2017). Cereal unit has been applied as the functional unit in at least one other LCA of crop rotations, is used in some national agricultural statistics and accounts for the most important nutritional functions of crops (Brankatschk and Finkbeiner 2014; Prechsl et al. 2017). It was therefore deemed appropriate for the present study. Drawbacks of using CU as the functional unit are that the values are based on animal feeding value although not all crops are used for animal feed, and that the data on livestock species composition are based on German conditions. However, since more Swedish cereals are used for animal feed than for human consumption (Eklöf 2014), and since livestock species composition in Germany and Sweden are similar (FAO 2016), this compromise was considered acceptable in the present study. Charles et al. (2006) included a quality criterion in their LCA of wheat cultivation, which is relevant since fertiliser management affects protein content in harvested crops, but difficult to apply when assessing whole crop rotations, since different characteristics are valued in different crops. In addition, low-quality crops are not always discarded but instead used for other purposes, e.g. low-protein wheat can be used as animal feed or for biofuel production. We therefore chose not to include any crop quality criteria in this study, although it should be noted that the quality aspect should be considered when evaluating appropriate fertiliser management in a decision-making context.
Environmental impacts of agricultural systems are dominated by soil emissions, which in turn are site-dependent (Notarnicola et al. 2017). Since measured emission data are rarely available, the reliability of LCA outcomes depends on the availability of emission models that can be applied with existing data. Our sensitivity analysis showed that estimated GHG emissions were highly sensitive to changes in soil N2O emissions. Feasible models for estimating N2O emissions at field scale are scarce, mainly due to emissions being highly variable over time and space (Butterbach-Bahl and Dannenmann 2011), which is why the crude Tier I model presented by IPCC (De Klein et al. 2006) is used in most crop LCAs. The model used to estimate soil N2O emissions in this study was instead derived from a meta-analysis of Canadian field trials, and has not been verified for Swedish conditions. While this is an important source of uncertainty in the quantitative results, it should give a more realistic representation of differences in emissions between sites than using a site-independent emission factor. Similarly, marine eutrophication was highly sensitive to changes in nitrate leaching, since this emission dominated the impacts assessed. The model used for estimating nitrate leaching is based on data derived from national reporting, together with correction factors for fertiliser rates over or under optimum level originally developed for a national farmer’s advisory tool in Sweden. Considering the large contribution of soil N2O and nitrate emissions to these two important impact categories, harmonised models that are globally applicable but still account for spatial and management variations would improve both the accuracy and inter-comparability of crop LCAs.
The third largest contributor to the impacts assessed was SOC changes, which are uncertain both in terms of methodological choice (see Methods section) and in terms of empirical data (Fig. 4). The confidence intervals of both SOC change per ha (Fig. SM2) and resulting CO2 emissions per CU (Fig. 4) exhibited considerable overlap, which complicates interpretation of the results. Modelling SOC dynamics instead of using measured values would be an option to achieve more stable results, but would also introduce new uncertainties associated with the chosen model and methodological choices such as temporal system boundary and choice of initial SOC level. Measured data were therefore chosen as the least biased way to represent results, in accordance with recommendations by Goglio et al. (2015), but the large confidence intervals mean that interpretation of observed impacts due to SOC changes is uncertain.
Implications and perspectives
While the medium N level had the lowest GHG emissions and marine eutrophication impact, it required more land per unit produced than the high N level (Figs. 3 and 6). Agricultural land is a limited resource, so reducing the area of land required for producing a unit of crops can potentially prevent land use change, or free up space for environmental impact mitigation measures. Examples of these mitigation measures are constructed wetlands to retain nutrients (Land et al. 2016) and producing biomass that promotes soil carbon sequestration and replaces fossil fuels (Hammar et al. 2014; Prade et al. 2014). Several recent studies have indicated high climate mitigation potential of reducing land requirements through agricultural intensification (Balmford et al. 2018; Searchinger et al. 2018), while others claim that yield improvements do not necessarily mean that less cropland is actually used (Lambin and Meyfroidt 2011). These discrepancies highlight the importance of considering the system beyond the field scale in order to make a fair assessment of the environmental consequences of different management practices. However, the results presented in this study show that the field-level environmental performance response to different fertiliser intensities is site-dependent, which is potentially also the case for other proposed intensification measures, such as altering tillage practices or introducing catch crops (see e.g. Doltra and Olesen 2013; Zaher et al. 2013). The highly site-dependent nature of agricultural systems is therefore relevant to consider when evaluating the mitigation potential of intensification strategies at larger scales.
Site-dependent effects of management practices and management change have been reported previously (Goglio et al. 2014). The difference in impact magnitude and preferred N level for minimising impacts at sites located geographically close to each other in this study (Figs. 1, 3 and 5), as well as other patterns in the results, indicate that soil texture was one of the most important variables. This outcome indicates possibilities for decreasing the environmental impact by considering soil characteristics when planning e.g. crop rotations and fertiliser strategy, as farmers already do to maximise profit. Since soil texture can vary even within fields, it is also possible that precision fertilisation could have a significant effect on the overall environmental impact (Delin and Stenberg 2014). However, further research is needed to identify factors that can be used to characterise sites in terms of their environmental impact profile. The variations in environmental impact between sites in this study illustrate how using emissions and impact assessment models operating at a relevant spatial scale in relation to the research question improves the possibility of drawing relevant conclusions.
This study explored the influence of site and N fertiliser rate on the GHG emissions and marine eutrophication impact from crop cultivation in a life cycle perspective. Results from a 20-year assessment at nine sites and three N fertiliser levels revealed that site affected the N level that gave the lowest impacts and the impact level in general, and that results were also not consistent between impact categories. This outcome illustrates that general management plans for decreasing the environmental impact of crop cultivation will have difficulty succeeding without considering site characteristics and potential trade-offs between different impacts.
Overall, the results showed that site influenced GHG emissions and marine eutrophication more than N level did, at least for the moderate fertiliser rates studied here. The medium N level, which was lower than current average rate in the study regions, gave on average the lowest total GHG emissions and marine eutrophication. However, differences between mean impacts at each N level were small (up to 10% for GHG emissions and 20% for marine eutrophication) and not statistically significant for total GHG emissions, and only significant between medium and high N level for marine eutrophication. In contrast to the moderate differences observed between N levels, differences between mean impacts at the different sites were large (up to 280% for GHG emissions and 340% for marine eutrophication) and statistically significant for both impact categories. These results show that site-specific information can improve the accuracy of assessments of the environmental impact of crop cultivation and thereby generate better decision support.
The authors would like to thank Claudia von Brömssen at the Department of Energy and Technology, Swedish University of Agricultural Sciences, for providing guidance on the statistical analyses. Agricultural Sciences and Spatial Planning (Formas) [Grant number 229-2013-82]. This work was funded by the Swedish Research Council for Environment.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
- Ahlgren S, Hansson P-A, Kimming M, Aronsson P, Lundkvist H (2009) Greenhouse gas emissions from cultivation of agricultural crops for biofuels and production of biogas from manure—implementation of the Directive of the European Parliament and of the Council on the promotion of the use of energy from renewable sources. In: Revised according to instructions for interpretation of the Directive from the European CommissionGoogle Scholar
- Aoun WB, El-Akkari M, Gabrielle B, Flenet F (2014) Considering the variability of farming practices improves the LCA of biodiesel from oilseed rape. In: Proceedings of the 9th international conference on life cycle assessment in the agri-food sector. American Center for Life Cycle Assessment, San FranciscoGoogle Scholar
- Aronsson H, Torstensson G (2004) Beräkning av olika odlingsåtgärders inverkan på kväveutlakningen. Beskrivning av ett pedagogiskt verktyg för beräkning av kväveutlakning från enskilda fält och gårdar [Swedish]. Swedish University of Agricultural Sciences, UppsalaGoogle Scholar
- Ashworth AJ, Taylor AM, Reed DL, Allen FL, Keyser PD, Tyler DD (2015) Environmental impact assessment of regional switchgrass feedstock production comparing nitrogen input scenarios and legume-intercropping systems. J Clean Prod 87:227–234. https://doi.org/10.1016/j.jclepro.2014.10.002 CrossRefGoogle Scholar
- Brander M, Sood A, Wylie C, Haughton A, Lovell J (2011) Electricity-specific emission factors for grid electricity. Econometrica. https://ecometrica.com/assets/Electricity-specific-emissionfactors-for-grid-electricity.pdf. Accessed 29 Oct 2018
- Brentrup F, Kusters J, Lammel J, Barraclough P, Kuhlmann H (2004) Environmental impact assessment of agricultural production systems using the life cycle assessment (LCA) methodology—II. The application to N fertilizer use in winter wheat production systems. Eur J Agron 20:265–279. https://doi.org/10.1016/s1161-0301(03)00039-x CrossRefGoogle Scholar
- Brentrup F, Hoxha A, Christensen B (2016) Carbon footprint analysis of mineral fertilizer production in Europe and other world regions. In: Paper presented at the LCA Food 2016, DublinGoogle Scholar
- De Klein C, Novoa RSA, Ogle S, Smith KA, Rochette P, Wirth TC, McConkey BG, Mosier A, Rypdal K, Walsh M, Williams SA (2006) Chapter 11: N2O emissions from managed soils, and CO2 emissions from lime and urea application. In: Gytarsky M, Hiraishi T, Irving W, Krug T, Penman J (eds) 2006 IPCC guidelines for national greenhouse gas inventories. IPCC International Panel on Climate Change, Geneva, pp 11.1–11.54Google Scholar
- Edström M, Pettersson O, Nilsson L, Hörndahl T (2005) Jordbrukssektorns energianvändning [Swedish]. Swedish Institute of Agricultural and Environmental Engineering (JTI), UppsalaGoogle Scholar
- Eklöf P (2014) Marknadsöversikt – Spannmål. Jordbruksverket (Swedish Board of Agriculture)Google Scholar
- EMEP/EEA (2016) EMEP/EEA air pollutant emission inventory guidebook 2016. In: Technical guidance to prepare national emission inventories, vol EEA report no 21/2016. European Environment Agency, LuxembourgGoogle Scholar
- European Commission (2017) Modernising and simplifying the cap. In: Background document: climate and environmental challenges facing EU agriculture and rural areas. European Commission. Directorate-General for Agriculture and Rural Development, BrusselsGoogle Scholar
- FAO (2016) FAOSTAT statistics database: live animals, by 1058 country. http://www.fao.org/faostat/en/#data/QA. Accessed 22 Feb 2019
- Fertilizers Europe (2018a) Ammonium nitrate (AN, 33.5% N); from ammonia and nitric acid, including primary production; production mix, at plant. UUID: c2c4ebba-358f-493e-83ba-71bffe847e9a (version 00.00.001). thinkstep, GaBi databaseGoogle Scholar
- Fertilizers Europe (2018b) Potassium chloride (KCl/MOP, 60% K2O); shaft mining and beneficiation, including primary production; production mix, at plant. UUID: a2a8695e-968c-4341-922c-a007c8a8c56d (version: 00.00.001). thinkstep, GaBi databaseGoogle Scholar
- Fertilizers Europe (2018c) Triple superphosphate (TSP, 46% P2O5); rock phosphate acidulation with phosphoric acid, including primary production; production mix, at plant. UUID: 8d0007f0-9ad8-43b0-86e4-ebe6e9f9d0e6 (version 00.00.001). thinkstep, GaBi databaseGoogle Scholar
- Gode J, Martinsson F, Hagberg L, Öman A, Höglund J, Palm D (2011) Miljöfaktaboken 2011. Estimated emission factors for fuels, electricity, heat and transport in sweden. Värmeforsk, StockholmGoogle Scholar
- Huttunen M, Vehviläinen B, Huttunen I (2013) Typen, fosforin ja kiintoaineksen pidättyminen vesistöissä–WSFS-Vemala-mallin arvio. Finnish Environment Institute, HelsinkiGoogle Scholar
- ISO (2006) ISO 14040: environmental management-life cycle assessment-principles and framework, 2. utg. edn. International Organization for Standardization (ISO), GenevaGoogle Scholar
- Johnsson H, Mårtensson K, Lindsjö A, Persson K, Rangel YA, Blombäck K (2016) Läckage av näringsämnen från svensk åkermark. Beräkningar av normalläckage av kväve och fosfor för 2013 vol SMED Rapport Nr 189 2016. Sveriges Meteorologiska och Hydrologiska Institut, NorrköpingGoogle Scholar
- Licker R, Johnston M, Foley JA, Barford C, Kucharik CJ, Monfreda C, Ramankutty N (2010) Mind the gap: how do climate and agricultural management explain the ‘yield gap’ of croplands around the world? Glob Ecol Biogeogr 19:769–782. https://doi.org/10.1111/j.1466-8238.2010.00563.x CrossRefGoogle Scholar
- Lindgren M, Pettersson O, Hansson P-A, Norén O (2002) Engine load pattern and engine exhaust gas emissions from off-road vehicles and methods to reduce fuel-consumption and engine exhaust gas emissions [Swedish]. Swedish Institute of Agricultural and Environmental Engineering (JTI), UppsalaGoogle Scholar
- Myhre G et al (2013) Anthropogenic and natural radiative forcing. In: Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM (eds) Climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change. Cambridge University Press, CambridgeGoogle Scholar
- Nemecek T, Schnetzer J (2011a) Data collection of inputs and yields in LCIs of agricultural production systems in Switzerland and other European countries. Data v3.0. Agroscope, ZürichGoogle Scholar
- Nemecek T, Schnetzer J (2011b) Data collection of inputs and yields in LCIs of agricultural seed production systems and seed processing-Data v3.0. Agroscope, ZürichGoogle Scholar
- Peter C, Fiore A, Hagemann U, Nendel C, Xiloyannis C (2016) Improving the accounting of field emissions in the carbon footprint of agricultural products: a comparison of default IPCC methods with readily available medium-effort modeling approaches. Int J Life Cycle Assess 21:791–805. https://doi.org/10.1007/s11367-016-1056-2 CrossRefGoogle Scholar
- Prade T, Svensson SE, Björnsson L (2014) Introduction of grass-clover crops as biogas feedstock in cereal-dominated crop rotations. Part I: effects on soil organic carbon and food production. In: Schenck R, Huizenga D (eds) Proceedings of the 9th international conference on life cycle assessment in the agri-food sector (LCA Food 2014), 8–10 October 2014, San Francisco. ACLCA, Vashon, pp 1032–1041Google Scholar
- Prechsl UE, Wittwer R, van der Heijden MGA, Lüscher G, Jeanneret P, Nemecek T (2017) Assessing the environmental impacts of cropping systems and cover crops: life cycle assessment of FAST, a long-term arable farming field experiment. Agric Syst 157:39–50. https://doi.org/10.1016/j.agsy.2017.06.011 CrossRefGoogle Scholar
- Statistics Sweden (2011) Växtskyddsmedel i jord-och trädgårdsbruket 2010. Användning i grödor. [Swedish]. Plant protection products in agriculture and horticulture. Use in crops. MI 31 SM 1101, Statistiska centralbyrån (SCB)Google Scholar
- Statistics Sweden (2016) SCB statistikdatabasen: [Elektronic resource] Statistical database. Statistiska centralbyrån, StockholmGoogle Scholar
- Struijs J, Beusen A, van Jaarsveld H, Huijbregts MAJ (2013) Eutrophication. In: Goedkoop M, Heijungs R, Huijbregts M, De Schryver A, Struijs J, van Zelm R (eds) ReCiPe 2008. Report I: characterisation. First edition (version 1.08)Google Scholar
- Sutton MA et al (2013) Our nutrient world: the challenge to produce more food and energy with less pollution. Centre for Ecology and Hydrology, Edinburgh on behalf of the Global Partnership on Nutrient Management and the International Nitrogen InitiativeGoogle Scholar
- Swedish EPA (2006) Eutrophication of Swedish seas: final report. Report 5509. Swedish Environmental Protection Agency, StockholmGoogle Scholar
- Swedish Meteorological and Hydrological Institute (2018) SMHI Öppna data. Meteorologiska observationer. https://opendata-download-metobs.smhi.se/explore/. Accessed 16 May 2018
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.