• Open Access

The use of life-cycle assessment to evaluate the environmental impacts of growing genetically modified, nitrogen use-efficient canola


* Correspondence (fax 44 118 9352421; e-mail j.r.park@rdg.ac.uk)


Agriculture, particularly intensive crop production, makes a significant contribution to environmental pollution. A variety of canola (Brassica napus) has been genetically modified to enhance nitrogen use efficiency, effectively reducing the amount of fertilizer required for crop production. A partial life-cycle assessment adapted to crop production was used to assess the potential environmental impacts of growing genetically modified, nitrogen use-efficient (GMNUE) canola in North Dakota and Minnesota compared with a conventionally bred control variety. The analysis took into account the entire production system used to produce 1 tonne of canola. This comprised raw material extraction, processing and transportation, as well as all agricultural field operations. All emissions associated with the production of 1 tonne of canola were listed, aggregated and weighted in order to calculate the level of environmental impact. The findings show that there are a range of potential environmental benefits associated with growing GMNUE canola. These include reduced impacts on global warming, freshwater ecotoxicity, eutrophication and acidification. Given the large areas of canola grown in North America and, in particular, Canada, as well as the wide acceptance of genetically modified varieties in this area, there is the potential for GMNUE canola to reduce pollution from agriculture, with the largest reductions predicted to be in greenhouse gases and diffuse water pollution.


Nitrogen is a primary nutrient for crops, and its usage has grown rapidly over the last 50 years. Current global nitrogen fertilizer production is in the region of 100 million tonnes, which takes about 1% of primary energy to produce. In the USA, about 13 million tonnes of ammonia are used for fertilizer, the production of which uses about 4.5 × 108 GJ of energy (World Energy Council, 2006), with resultant greenhouse gas emissions. In relation to climate change, the global warming potential (GWP) of N2O is approximately 300 times greater than that of carbon dioxide (CO2). Williams et al. (2006) reported that N2O from the nitrogen cycle dominates GWP from field crops, contributing about 80%. Globally, the agricultural sector is the second largest contributor of greenhouse gas emissions, at 17% of global emissions, vs. 31% for electrical and heat generation and 15% for transportation [Climate Analysis Indicators Tool (CAIT) Version 4.0, World Resources Institute, Washington DC, 2006]. Nitrogen fertilizer is estimated to account for approximately one-third of total greenhouse gas emissions from agriculture. Furthermore, a large proportion of applied nitrogen is not taken up by crops, often less than 50% (Addiscott et al., 1991). Although nitrogen not taken up by growing plants can become bound within organic matrices within the soil, significant amounts can be lost from the soil in gaseous form or leached in water (Wild, 1988). Potential contamination of groundwater by nitrates, which causes eutrophication and the pollution of drinking water, is also a major concern in many parts of the world (Kidd, 2002).

Recently, genetically modified, nitrogen use-efficient (GMNUE) canola has been developed using a Hordeum vulgare gene that enables greater efficiency of nitrogen uptake and use within the plant (see Arcadia, 2006). In effect, this means that the yield can be maintained with a substantially lower nitrogen fertilizer application. This article presents the methodology used and the results obtained from a life-cycle assessment specifically adapted to cropping systems for the production of 1 tonne of canola from a GMNUE variety compared with a conventionally bred control variety. The analysis was based on data from crop growth trials carried out in North Dakota and Minnesota in the seasons 2003 and 2004 in which nitrogen (applied as urea containing 46% nitrogen) was applied at rates of between 56 and 224 kg/ha. The nitrogen application rates and crop yields are shown in Table 1.

Table 1.  Yields of genetically modified nitrogen use-efficient (GMNUE) and conventional (Conv) canola at different nitrogen levels
SiteLevel of residual N in soil(NO3-N 60 cm/ha)*Nitrogen application kg/haYield of GMNUE canola kg/haYield of Conv canola kg/haPost-harvest N in soil GMNUE(NO3-N 60 cm/ha)Post-harvest N in soil Conv(NO3-N 60 cm/ha)Difference between yields kg/ha
  • nd, not detected.

  • *

    NO3-N 60 cm/ha is the amount of nitrate nitrogen in the top 60 cm of soil.

Minnesota 2003
 300  036592629ndnd1030
   5635552919   636
  11236102927   683
  16835413151   390
  22437303488   242
North Dakota 2004 
 107.25  02458169316.015.2 765
   562904264714.415.2 257
  1123328279515.215.2 533
  1683745298820.016.8 757
  2243698331120.820.8 387

The assumption is made that the production results obtained from replicated field-scale trials could be transferred to farm level if farmers had access to this new technology. Currently, about 5.5 million hectares of canola is grown in Canada, usually spring sown, requiring 90–110 days to maturity. The yields normally range from 1.5 to 2 tonnes/ha, with average rates of nitrogen fertilizer application of 80 kg/ha.

Life-cycle assessment

Life-cycle assessment is a recognized analytical method for ‘evaluating the environmental burdens associated with a product, process or activity by identifying and quantifying the energy and materials used and wastes released to the environment’[Society of Environmental Toxicology and Chemistry (SETAC), 1993]. It is used to assess the environmental impact of the production of a product, from raw material extraction through to final waste disposal and decomposition. In this respect, it is often termed a ‘cradle to grave’ technique. A life-cycle assessment will usually consist of four main phases that run sequentially: goal and scope definition; inventory analysis; impact assessment; and interpretation [SETAC, 1993; Viaammse Instelling voor Technologisch Onderzoek (VITO), 1995]. A comprehensive review of life-cycle assessment and the wide range of areas of application can be found in the European Environment Agency's (1997) Environmental Issues Series No. 6.

In the current study, the accepted guidelines [International Organization for Standardization (ISO), 1998] were used to undertake a life-cycle assessment specifically adapted to crop production to estimate the environmental burdens associated with growing GMNUE and conventional canola. These environmental burdens encompass a wide range of emissions.


Description of the canola production system

The canola production system analysed in this study was based on the cropping results from two seasons: Minnesota in 2003 and North Dakota in 2004. Plant protection agents were applied to each plot equally and field operations were carried out equally on each plot. The plots measured 900 m2 and were all spring-sown canola varieties. The conventional variety used was Brassica napus cv. Westar and the GMNUE line was BNAT-043, whose background variety is cv. Westar. All plots were hand weeded. The experimental recordings were bootstrapped (in Excel) to compensate for the small data sample, so as to provide an estimate of sample distribution (see Efron and Gong, 1983).

Description of the life-cycle assessment methodology

The life-cycle assessment methodology used in the current study followed that set by SETAC (1993). There are four interrelated stages to the methodology, and a brief outline of each stage is presented below.

Goal and scope definition

Goal and scope definition describes the aims, objectives and extent of the study. This study investigated two farming systems in which the function was to produce 1 tonne of canola. The research concentrated on those aspects of production for canola that were different from conventional systems and from GMNUE canola systems. Thus, the analysis focused on the quantities of nitrogen applied to the crop and the resulting yields. The ‘farm gate’ was used as the system boundary (see Figure 1). The study defines a functional unit which has a dual purpose, as it not only defines the unit to be assessed, but also reflects the product's utility. In this research, the functional unit was defined as ‘1 tonne of canola seed at the farm gate’.

Figure 1.

A schematic diagram showing the processes involved in a partial life-cycle assessment.

Inventory analysis

The inventory analysis compiles all resources required and all emissions released by the system under investigation and relates them to the defined functional unit (ISO, 1998). Production systems for both the conventional and GMNUE canola crops were modelled using data from Ecoinvent Centre (2004). This was substantiated by data on canola growing systems (Canola Council of Canada, 2006) and on emissions associated with field operations (Audsley et al., 1997). The analysis includes the manufacture, packaging and transport of nitrogen to the farm, in the form of urea; field operations; manufacturing, repair and maintenance of farm machinery; production and use of fuels for tractors; transport costs; and construction, maintenance and ultimate demolition of buildings for machinery storage. It also takes into account the emissions to air from combustion and emissions to soil from tyre abrasion during the work process.

Impact assessment

The emissions from each cropping system were characterized and classified to indicate the potential effects on a number of environmental impact factors that accumulate from the production of 1 tonne of canola.

The impact assessment stage of a life-cycle assessment characterizes and assesses the effects of the environmental burdens or emissions identified and quantified in the inventory. The emissions are grouped into environmental loadings (classification). These loadings are then multiplied by characterization factors to give a single valuation to facilitate comparison between environmental impact categories:

Impact category indicatori = Σj(Ej or Rj) × CFi,j(1)

where impact category indicatori is the indicator value per functional unit for impact category i, Ej or Rj is the release of emission j or consumption of resource j per functional unit and CFi,j is the characterization factor for emission j or resource j contributing to impact category i.

The characterization factors represent the potential of a single emission or resource consumption to contribute to the respective impact category (ISO, 2000). Such categories include GWP, acidification potential, eutrophication, human toxicity, and freshwater and marine ecotoxicity (ISO, 2000). A variety of different substances can contribute to each impact category, and one substance can contribute to more than one impact category.

In this life-cycle assessment, two baseline characterization methods, developed by Guinée (2002), were used for the main impact categories: these are characterizations of the Intergovernmental Panel on Climate Change (IPCC) (IPCC, 2001) and the Centre of Environmental Science of Leiden, known as CML 2001 (Guinée, 2002).

IPCC characterizations. IPCC (2001) describes how the potential for global warming for different gaseous emissions is characterized according to their GWP, and then aggregated into the impact category climate change. GWP is an index for estimating the relative global warming contribution caused by the atmospheric emission of 1 kg of a particular greenhouse gas in comparison with the emission of 1 kg of CO2. GWP is meant to compare emissions of long-lived, well-mixed gases, such as CO2, CH4, N2O and hydrofluorocarbons.

CML 2001 characterizations.  CML 2001 describes a ‘problem orientated approach’ which results in a categorization of indicators in relation to impacts such as eutrophication, acidification and ecotoxicology. This numerical representation of the level of an impact category, such as ecotoxicology, refers to the impact of toxic substances on ecosystems or human health. The ecotoxicity potential is calculated for each emission of a toxic substance to air, water and soil in kilograms of 1,4-dichlorobenzene (1,4-DCB) equivalent per kilogram of emissions.


The interpretation phase consists of an evaluation of the life-cycle assessment method and the results of the analysis, particularly in terms of the rigour and robustness of both the method and analysis, and conclusions can then be drawn.


The trial results presented in Table 1 illustrate the ability of GMNUE canola to take up and utilize soil-available nitrogen effectively. In Minnesota, the results illustrate little yield response of the GMNUE variety to increasing amounts of applied nitrogen. However, the yield of GMNUE canola under zero nitrogen application is equal to or greater than the yield of conventional canola under the highest nitrogen level. This is probably a result of the more effective use of the existing soil nitrogen reserves to accumulate biomass, when compared with the conventionally bred control variety. Figure 2 illustrates the energy consumption in MJ/tonne of canola grain at each of the fertilizer applications. This energy consumption can be categorized as follows:

Figure 2.

The energy consumption in MJ/tonne of canola grain for each of the fertilizer applications. The yield values are also included in the figure in order to illustrate the different productivity of the treatments. The error bars show the 90% confidence intervals derived from bootstrapping the yield data. GM, genetically modified, nitrogen use-efficient canola; Conv, conventionally bred control canola.

  • 1Production and delivery energy: the energy input into the processes which extract, process, refine and deliver energy or material inputs to a process.
  • 2Process energy: energy input required and consumed by the considered process to operate within the process phase, excluding production and delivery energy (for example, diesel consumption by a tractor).
  • 3Inherent energy: extracted energy which remains in the product after its production and delivery to its site of use (for example, natural gas in the production of mineral fertilizers).

The yield values are also included in Figure 2 to illustrate the different productivity of the treatments and to allow ready comparison of energy requirements between GMNUE and conventional varieties under equivalent yield conditions. For example, the yield of GMNUE canola at the 56 kg/ha nitrogen application was equal to or greater than that of conventional canola at the 112 kg/ha nitrogen application, yet a third less total energy was required.

The full range of characterized emissions is summed under each category in order to provide total values that can be expressed as a range of environmental burdens. Table 2 gives an example of the range of emissions which show the potential of the agricultural system to contribute to global warming or climate change. Figures 3 and 4 compare the emissions from the production systems in relation to GWP, eutrophication and acidification across both North Dakota and Minnesota field trial sites.

Table 2.  Levels of emission (kg CO2 equivalents) for both crop scenarios that in combination give the potential for global warming
Amount of nitrogen applied (kg/ha)kg CO2 equivalentGMNUE canolaConventional canola
North Dakota
Total carbon dioxide59.4491.88116.47135.77170.5786.27100.82138.68170.18190.06
Total carbon monoxide0.800.710.660.620.661.160.780.780.770.739
Total methane5.368.2110.3712.0715.157.789.0112.3515.1316.88
Total ethane0.020.020.0270.0280.
Total dinitrogen monoxide0.520.720.881.001.240.760.791.051.251.38
Total emissions66.16101.57128.42149.50187.6696.02111.45152.92187.39209.11
Total carbon dioxide37.3172.38104.73140.89166.1451.9288.15129.17158.30177.65
Total carbon monoxide0.500.550.570.620.620.700.670.710.690.66
Total methane3.366.469.3212.5214.754.687.8711.5014.0715.78
Total ethane0.
Total dinitrogen monoxide0.320.570.791.031.200.450.690.971.161.29
Total emissions41.5279.99115.45155.11182.7657.7897.43142.39174.28195.42
Figure 3.

Global warming potential (kg CO2 equivalent) (a), acidification (kg SO2 equivalent) (b) and eutrophication (kg NOx equivalent) (c) associated with canola production in North Dakota and Minnesota when using conventionally bred control canola (Conv) compared with genetically modified nitrogen use-efficient canola (GM). The error bars show the 90% confidence intervals derived from bootstrapping the yield data.

Figure 4.

Freshwater (a) and terrestrial (b) ecotoxicity [kg 1,4-dichlorobenzene (1,4-DCB) equivalent] associated with canola production in North Dakota and Minnesota when using conventionally bred control canola (Conv) compared with genetically modified nitrogen use-efficient canola (GM). The error bars show the 90% confidence intervals derived from bootstrapping the yield data.

Climate change

GWP is used to express the contribution made by gaseous emissions from all manufacturing systems to the problem of climate change. This is expressed as CO2 equivalents. A mean across all scenarios showed that the production of 1 tonne of GMNUE canola resulted in the emission of 114 kg CO2 equivalents, in comparison with 137 kg CO2 equivalents for the conventionally bred control variety. There were small differences between the North Dakota and Minnesota trials, with the latter having slightly lower overall emissions in terms of CO2 equivalents (104 vs. 121 kg CO2 equivalents). Figure 3a shows that the emission (kg CO2 equivalents) increases as more nitrogen is applied. As a mean across all nitrogen application regimes, the conventionally bred control variety emitted 16.3% more greenhouse gases per unit of crop produced than did the GMNUE crop. The mean value may underestimate the contribution of the GMNUE trait in practice. For example, the yield of GMNUE canola grown under 56 kg nitrogen in North Dakota was equal to or greater than the yield of the conventional variety grown under 112 kg nitrogen, yet 33% lower total emissions were produced.


The acidification potential of a system, expressed as SO2 equivalents, represents its contribution to the acidification of natural ecosystems, such as forests and lakes. Figure 3b relates to acidification and shows that the emissions (kg SO2 equivalents) increase as more nitrogen is applied. However, taking a mean across the different scenarios, GMNUE canola reduced acidification by 16.2%. Again, the mean value may underestimate the contribution of this trait in practice. For example, the yield of GMNUE canola grown under 56 kg nitrogen in North Dakota was equal to or greater than the yield of the conventional variety grown under 112 kg nitrogen, yet 33% lower kg SO2 emissions were produced.


Terrestrial eutrophication is caused by the atmospheric deposition of nutrients on natural land systems. It is expressed as kg NOx equivalents per tonne of grain. Direct aquatic eutrophication is not stated here, as the effects from manufacture and application are diffuse and very dependent on local edaphic factors, such as soil type and structure, and the proximity of water bodies. However, terrestrial eutrophication assumes that a proportion of the total figure leaches through to the groundwater. Figure 3c illustrates the potential for eutrophication through emissions of kg NOx equivalents. As a mean across the scenarios, the conventionally bred control variety emitted 18% more kg NOx equivalents per unit of crop produced than did the GMNUE crop. In this case, eutrophication is associated with the manufacture of machinery, transport and field operations, and not purely with the direct application of fertilizer.


This category is used to describe the effect of any toxic emissions produced during the manufacturing process on terrestrial and aquatic ecosystems. These are expressed as kg 1,4-DCB equivalents. Figure 4 illustrates the emissions for aquatic and terrestrial ecotoxicity levels across both field trial sites. Taking a mean across the scenarios, there are benefits from GMNUE canola in terms of aquatic and terrestrial ecotoxicity reductions of 17.3% and 15.4%, respectively. Specifically, GMNUE canola grown under 56 kg nitrogen in North Dakota produced a yield equal to or greater than the yield of the conventional variety grown under 112 kg nitrogen, with approximately 25% and 40% less potential for aquatic and terrestrial ecotoxicity, respectively.

In each case, the GMNUE canola trials produced markedly lower levels of each type of emission in comparison with conventional systems.


The potential impact of any genetically modified crop on the environment is a key issue (Bennett et al., 2004, 2006a,b). Increasing concern about climate change (for instance, see Stern Review, 2006) means that the potential for genetically modified crops to cut greenhouse gas emissions requires more detailed analysis.

The analysis of the production of canola in these two different locations illustrates that growing GMNUE canola leads to an energy saving of 22% when compared with the conventionally bred variety. This relates to the lower nitrogen levels required to achieve similar yields using GMNUE canola. If the current 5.5 million hectares of canola grown in Canada were planted with the GMNUE variety, this would result in a saving of approximately 1.5 × 106 GJ of energy, or 250 000 barrels of oil (assuming that one barrel of oil contains 6.1 GJ of energy). This figure is likely to underestimate the true potential impact of this technology, as it is based on an average across all conditions.

Based on the current 8.5 million tonnes of canola grown in Canada, and assuming similar nitrogen use efficiency gains, conversion to GMNUE canola in Canada would reduce emissions by 170 000 tonnes of CO2 equivalents. The introduction of the GMNUE trait in countries such as China and India, which produce about 40% of the world's rapeseed, has the potential to significantly reduce energy use and thus greenhouse gas production. About 50 million tonnes of canola are grown worldwide, and conversion to GMNUE canola could reduce emissions by 1 million tonnes of CO2 equivalents.

The possible decrease in the effects of diffuse agricultural pollution is of particular relevance in North America where, under the National Environmental Policy Act of 1969, the American government has the responsibility to ‘attain the widest range of beneficial uses of the environment without degradation, or other undesirable and unintended consequences’; moreover, the Clean Water Act 33 USC § 1251 et seq. also aims to reduce the discharge of agricultural pollutants. If these crops were to be grown in Europe, however, the potential for reduced levels of water pollutants from agriculture has a particular relevance to the Water Framework Directive (WFD) 2000/60/EC (Commission of the European Communities, 2000). For example, in the UK, agriculture is seen as one of the main sources of the diffuse pollution of water. The life-cycle assessment reported here shows that GMNUE canola can produce yields equivalent to conventional varieties using significantly less fertilizer. The resultant lower applications of fertilizer reduce the risk of nitrogen polluting water, thus decreasing watercourse enrichment, and could potentially help to meet targets set as part of the WFD.

Evidence from the uptake of previous genetically modified crops in Canada suggests that farmers recognize the benefits of growing these crops and will be ready to utilize GMNUE canola once released on to the market. For instance, the Canola Council of Canada has reported that a recent study on the agronomic and economic impact of herbicide-resistant transgenic canola varieties showed benefit to the grower, the industry and the environment. These canola varieties have shown a 10% increase in yield and a reduced number of tillage and herbicide passes in comparison with conventional varieties. Overall, growers of herbicide-resistant canola in Canada applied 6000 tonnes less chemicals and used 31 million litres less fuel, saving themselves $13 million (Canola Council of Canada, 2003).

Although the results of this study point to a number of environmental benefits associated with the cultivation of GMNUE canola, several points require further consideration. First, the analysis is based on a relatively few plot-scale trials and, although these suggest significant benefits, the scaled-up calculations should be regarded as estimates. To overcome some of these data shortfalls, bootstrapping was used as a relatively straightforward method for deriving errors and confidence intervals. Bearing in mind this estimation, our calculations suggest that, even if wide-scale field adoption did lead to a dilution of effects, these effects would still be significant and economic. The upward trends in oil prices, at the time of writing $75 a barrel, are likely to lead to further increases in the costs of nitrogen fertilizer and further enhance the profitability of growing GMNUE canola. Second, the article does not deal at all with any knock-on environmental impacts concerning the competitive ability of weed communities in relation to broad-acre canola production. From a purely productive perspective, any increase in the competitive ability of canola in low-nitrogen conditions may be regarded as beneficial. However, from the perspective of arable land biodiversity, the continued loss of arable weed species has been a cause for concern in some counties for a number of years (for instance, see Wilson et al., 1990). Finally, it must be noted that life-cycle assessment has certain limitations. The accuracy of a life-cycle assessment is dependent on the quality of the available inventory data used to produce the environmental impact effect. Guinée (2002: 8–9) notes in the context of life-cycle assessment that, ‘in practice, data are frequently obsolete, incomparable or of unknown quality’. In the current life-cycle assessment, the best data available were used and were consistent across the systems compared in the analysis.

Assuming that similar gains in nitrogen use efficiency can be made, it is clear from the analysis that there is the potential for GMNUE technology to be used more broadly in crops to reduce pollution from agriculture, particularly in terms of reducing energy use, the emission of greenhouse gases and diffuse water pollution. The results also show that there is the potential to apply life-cycle assessment methodology to compare and evaluate other cropping methodologies, such as minimal tillage, irrigation, fertilization systems, different rotations and comparative organic systems.


This work was supported by The University of Reading with additional financial support from Arcadia Biosciences, Inc. Field trial data were provided by Arcadia Biosciences, Inc. Any statements, errors or omissions in this paper remain the sole responsibility of the authors.