A Comparison of IPCC Guidelines and Allocation Methods to Estimate the Environmental Impact of Barley Production in the Basque Country through Life Cycle Assessment (LCA)

: This study aimed to estimate the environmental impact of barley production in the Basque Country, Northern Spain, using cradle-to-gate life cycle assessment (LCA) methodology, as well as to assess how methodological choices (i.e., the use of IPCC 2019 Guidelines versus allocation methods) can inﬂuence such estimation. The production of mineral fertiliser and the direct emissions of nitrous oxide (N 2 O) resulting from the application of nitrogen (N) fertiliser were identiﬁed as the two main contributors (40% and 30% of all greenhouse gas emissions, respectively) to the environmental impact of barley production. Pertaining to GHG emissions themselves, the use of calcium ammonium nitrate fertiliser was found to be the main contributor. Therefore, the optimization of N fertiliser application was established as a key process to reduce the environmental impact of barley production. The fertiliser-related release of N and phosphorous (P) to the environment was the main contributor to particulate matter formation, terrestrial acidiﬁcation, and terrestrial and marine eutrophication. The incorporation of environmental data on NH 3 , NO x , NO 3 − , and PO 43 − to the LCA led to a more accurate estimation of barley production impact. A sensitivity analysis showed that the use of economic allocation, compared to mass allocation, increased the estimation of climate change-related impact by 80%. In turn, the application of the IPCC 2019 Reﬁnement Guidelines increased this estimation by a factor of 1.12 and 0.86 in wet regions and decreased in dry regions, respectively. Our results emphasise the importance of the choice of methodology, adapted to the speciﬁc case under study, when estimating the environmental impact of food production systems.


Introduction
In the Basque Country (Northern Spain), barley (Hordeum vulgare L.) production makes up a large portion of the agricultural land use. Barley is a commonly cultivated feed and food crop, which is widely adaptable to unfavourable climate conditions and superior as a raw material for the malt and beer industries. Barley is the 12th most important agricultural product in the world, with Europe being the largest producer (60.2%) [1]. In Spain, according to recent Eurostat data, the annual production of barley is approximately 6.7 and 3.6 million tons of grain and straw, respectively [2]. In the Basque Country, barley production is located in the southern province of Araba/Álava, with an area of nearly 14,000 hectares and an annual production of 59,200 to 95,800 tonnes in the five-year period of 2015-2019 [3], making this region the largest producer of malting barley in Spain (10% of national production).
The estimation and assessment of the environmental impact of agricultural practices and systems is key to promote the sustainability of the agricultural sector. Understanding

Materials and Methods
The LCA methodology implemented in SimaPro 9.1.1.1 software was chosen to perform the abovementioned environmental analysis, according to the principles described in the corresponding ISO standards and considering the following four phases: (i) goal and scope definition; (ii) life cycle inventory (LCI); (iii) life cycle impact assessment (LCIA); and (iv) life cycle interpretation. Finally, a sensitivity analysis was carried out to determine the robustness of our LCA results and their sensitivity to uncertainty factors and to identify the most important set of model parameters to determine whether the data quality needs to be improved, and to enhance interpretation of results.

Goal and Scope Definition
This study was conducted to assess the environmental impact of barley production in the province of Araba/Álava, Basque Country, Northern Spain. In this respect, we aimed to compare the estimated environmental impact resulting from the application of different allocation methods (i.e., mass allocation vs. economic allocation) and the IPCC Guidelines (the IPCC 2006 Guidelines vs. the 2019 Refinement to the 2006 IPCC Guidelines, hereinafter called IPCC 2019) [16,17].
The results of this study are of interest to all those interested in the identification of the most important contributors ("critical hotspots") regarding the environmental sustainability of agricultural food production systems and, specifically, barley production.

Characteristics of the Study Area
The climate and soil conditions in the province of Araba/Álava are highly favourable for cereal cultivation (wheat, maize, rye, oat, and barley). In the five-year period of 2015-2019, an average of 13,900 ha was dedicated to barley cultivation, with an average yield of 5350 kg ha −1 (ranging from 4200 to 7000 kg ha −1 ).
The climate in the province of Araba/Álava is characterised by an average temperature of 11.6 • C (ranging from 10.2 • C to 13.2 • C). Peak temperatures are recorded during the summer. Rainfall averages 674 mm, ranging from 626 mm to 733 mm. The highest rainfall occurs in the autumn and winter seasons.
Taking into account the climatic data, at the municipality level, provided by SIGA (Agrarian Geographic Information System, Spanish Government) [15], as well as the IPCC 2019 guideline (i.e., wet climate (W) = ratio of annual precipitation to potential evapotranspiration >1; dry climate (D) = ratio of annual precipitation to potential evapotranspiration <1), the province of Araba/Álava can be disaggregated into two different climatic regions, wet (W) and dry (D), as shown in Figure 1.  Table A1 (Appendix A).

System Boundaries and Functional Unit
The system included all purchased inputs and activities carried out for barley production in the study area during the 2014-2015 and 2015-2016 growing seasons, following a cradle-to-gate approach from field preparation to crop harvest. Figure 2 shows the inventory data and the system boundaries taken into consideration in our estimation of the environmental impact of barley production in Araba/Álava, including the main inputs (i.e., seeds, fertilisers, pesticides, and fuel); the agricultural field operations (fuel consumption divided into three categories: soil preparation and sowing, crop management, and crop harvest); the outputs (barley grain and straw); and the emissions to air, soil, and water. The influence of transportation and the drying stage were not considered in our analysis.
The functional unit (FU), a quantified performance of a product system to be used as a reference unit in LCA, is the basis for comparisons between different systems [5]. Since the main function of the agricultural system under study here is barley grain production, the FU was defined as "one tonne of harvested barley grain (dry mass of 85%) at the farm gate", in agreement with related studies [10,12]. In addition to barley grain (the target product), straw was generated as a by-product.

Allocation Procedure
In this study, a mass-based allocation criterion was implemented as the first suggestion considered by the product category rules (PCR) for arable crops, which are defined as "sets of specific rules, requirements and guidelines for developing environmental declarations that use quantified environmental data" [18]. Permit analysis of the system that avoids market variations could occur from economic allocation. Mass allocation percentages were estimated by considering the biomass production (kg) of barley grain and barley straw, achieving the allocation factors of 49.50% and 50.50% for grain and straw, respectively. Data on barley straw production were not available for our studied plots. In consequence, barley straw production was estimated by considering a ratio of the above-ground residue dry matter to the harvested yield (R AG(T) ) of 1.2, according to the IPCC 2006 Guidelines [17].
For the sensitivity analysis (see below), the allocation procedure was based on economic criteria. According to Ardente and Cellura (2012), the economic allocation option should be considered, among other cases, whenever the prices of by-products and coservices differ widely [19]. For our economic allocation, annual average prices for the period of 2010-2020, available in databases from the Basque Government [20,21], were considered, thus assigning a greater percentage value for barley grain (89.29%) due to its higher price compared to straw. This percentage value is very similar to that proposed in the Ecoinvent Database [22], a compliant data source for studies and assessments, based on ISO 14040 and ISO 14044, reported in the PCR for arable crops. In fact, the Ecoinvent Database applies a value of 89.90% and 10.10% for grain and straw, respectively, when no other allocation criteria other than economic criteria can be applied.

Life Cycle Inventory (LCI)
Our life cycle inventory (LCI) was defined as follows: (i) seed and fertiliser production; (ii) pesticide production; (iii) agriculture field operations; and (iv) emissions from soil (N 2 O, NO x , NH 3 , NO 3 − and PO 4 3− ). Data for the LCI were obtained from farmer surveys carried out during the development of the LIFE AGROgestor [23] project. The corresponding farmers provided the required information on the most widely adopted agricultural practices used for barley production in the studied region. This information was incorporated to the LCA (Table A1, Appendix A). Table 1 shows a summary of the inventory flows (including the average quantities), as well as a brief description of the Ecoinvent processes, used in barley production in the studied region.
An average yield of 6.21 tonnes ha −1 of barley grain (dry mass = 85%) with a straw production of 6.33 tonnes ha −1 (dry mass = 85%) was established. We assumed that the straw (i) was not incorporated into the soil after crop harvesting and (ii) had an export rate of 90%. Following Noya et al. (2015) [4], in our assessment, no change in the overall soil carbon content was assumed because of the difficulty of its measurement, due to its dependency on a wide variety of factors such as management practices, climate, previous cropping regimes in the area, etc. Emission factors for the production of barley seed, fertiliser, and pesticide were taken from the Ecoinvent Database [24] (Table 1). Seed production was referenced to 1 kg of barley seed, with a maximum water content of 15%, at a regional storage centre. The seed rate used in our LCI was 228 kg ha −1 .
Since, in general, no organic fertilisers are used for barley production in the studied region, we assumed that only mineral fertilisers were applied with an average application rate of 158-80-24 kg of N-P 2 O 5 -K 2 O per hectare. The following mineral fertilisers were considered for our estimation: calcium ammonium nitrate (CAN: 400 kg ha −1 ) and NPK mixture (13-20-6: 400 kg ha −1 ). Since the emission factor for the specific NPK mixture used in the studied region (13-20-6) was not available, we used the factor for the NPK 15-15-15 mixture.
According to some authors [25], in terms of emissions, the contribution of pesticide application can be considered negligible. Nonetheless, emissions from pesticide use were estimated in this study. Since emission factors for each individual pesticide were not available, we assumed the same emission factor for all of them (Table 1), based on the active ingredients (a.i.) of unspecified pesticides. An average application of 0.48 kg a.i. ha −1 was estimated.
Transport and packaging of seeds, fertilisers, and pesticide were included in the system boundaries, as reflected in Table 1.

Emissions from Agricultural Field Operations (Fuel Consumption)
All field operations for barley cultivation in the studied region are normally carried out by the farmers themselves, including ploughing, harrowing, sowing, fertilisation, pesticide applications, and harvesting. Fuel consumption data for these agricultural practices were obtained from the above-mentioned surveys, including information on tractor power, machinery operation time (hours), and fuel consumption during field work activities (L ha −1 ). When consumption data were not available, they were estimated using Spanish databases (IDAE [22]) and/or the consumption values proposed by Lovarelli et al. (2020) [10]. According to the information obtained from the farmer surveys, as well as from these sources, diesel consumption was adapted to the study conditions, setting a specific value for each field operation (Table 1).
Emissions derived from fuel consumption for the different agricultural operations were obtained from the Ecoinvent Database (Table 1). The specific weight of diesel was set as 0.84 kg L −1 [25][26][27].

Emissions from Field
The accuracy, robustness, and reliability of LCA data depend on the quality and representativeness of the input-output data [14]. The inclusion of emission data, in addition to material purchases and energy flows, is a crucial stage of the LCA framework [5,28]. Therefore, emissions of NO x , NH 3 , NO 3 − , and PO 4 3− were included in our LCI due to its relevance for various impact categories, such as acidification and eutrophication [29]. Moreover, field emissions from fertilisation (N 2 O) affect GHG emissions [30] and, therefore, the climate change impact category.
Two different methodologies were considered to estimate the field emissions derived from mineral fertiliser application, as well as from crop residues, during barley production in the studied region: (i) the IPCC 2006 [17] methodology, as the baseline scenario, and (ii) the IPCC 2019 Refinement [16] methodology, as the sensitivity analysis (see below). Table 2 shows information on the N 2 O factors used here. Disaggregation of EF 1 by climate was calculated according to the IPCC 2019 Guidelines for wet (W) and dry (D) climates.
Pertaining to crop residues, we estimated that 90% of the straw is removed and, hence, only 10% remains as crop residue, being an additional source of nitrogen for nitrification and denitrification processes, leading to N 2 O emissions. The amount of N in crop residues (F CR ) was estimated from above-ground and below-ground crop residue biomass values multiplied by the corresponding N concentration, as defined by the IPCC Guidelines [16,17]. The emission factors for the N content of above-ground residues (N AG(T) ) and below-ground residues (N BG(T) ) were 0.007 and 0.014, respectively.  Nitrate (NO 3 − ) emissions due to leaching and runoff were estimated following the emission factors proposed by Zampori and Pant (2019) [31]. The emission values are reflected in the IPCC methodology, namely, 1.33 and 1.06 kg NO 3 − emitted per kg of N in fertilisers applied (IPCC 2006 [17] and IPCC 2019 [16], respectively). As indicated in the IPCC methodologies, this estimation only applies to wet climate regions.
Regarding ammonia emissions from fertiliser application, emission factors as reported in the EMEP/EEA Air Pollutant Emission Inventory Guidebook [32] were used, i.e., 17 and 91 g NH 3 kg −1 N applied for the CAN and NPK mixtures, respectively. For the selection of the NH 3 emission factor, a cool climate (i.e., average temperature below 15 • C) and a soil pH > 7.0 were assumed.
Emissions derived from fertiliser volatilisation (kg NO-N volatilised per kg of N in fertiliser) were differentiated according to the type of fertiliser, as per Zampori and Pant (2019) [31], as follows: 0.016 kg of NO-N for CAN and 0.0225 kg of NO-N for NPK mixtures as a mix of ammonium nitrate (50%, with a emission factor of 0.029 kg of NO-N per kg of N applied) and CAN (50%), in accordance with PCR 2020 [18]. Since no applications of lime, urea, or urea-containing compounds were made in the studied plots, their related emissions were not included.
Phosphate emissions (PO 4 3− ) were estimated following the Swiss agricultural life cycle analysis (SALCA) model [33], which distinguishes between the leaching of soluble phosphate to groundwater (P gw ) and the run-off of soluble phosphate to surface water (P sw ) that was corrected by mineral P fertilisation (F ro ) ( Table 3). Phosphorus correction factors of 0.07 kg P ha −1 year −1 and 0.175 kg P ha −1 year −1 were used for P leaching to groundwater in arable land and P lost through run-off to rivers, respectively [24,34].
Air, water, and soil emissions from pesticide application were estimated considering that 10, 8.5, and 76.5% of the pesticide active ingredient were released to the air, water, and soil, respectively. The remaining 5% of the active ingredient was assumed to be maintained on the crop itself, in line with recent research [35,36]. By contrast, more simplified approaches assume that pesticides are entirely released to the soil compartment, e.g., Ecoinvent [26]. For the simulation in SimaPro, all the pesticide active ingredients (a.i.) were considered to be glyphosate.
Except for N 2 O (Table 2), all the other emission factors used here are shown in Table 3. i.: pesticide active ingredient; P gw = quantity of P leached to groundwater (kg P ha −1 year −1 ); P gwl = average P leached to groundwater (0.07 kg P ha −1 year −1 for arable land); F gw = correction factor for fertilisation with slurry (no organic application in the present study); P ro = average P lost through run-off to rivers kg P ha −1 ; P rol = average quantity of P lost through run-off (0.175 kg P ha −1 year −1 for arable land and with slope ≥3%); F ro = correction factor for fertilisation with mineral and organic P applied (min = P 2 O 5 from mineral fertiliser, kg ha −1 ).

Life Cycle Impact Assessment (LCIA)
The environmental impact from the emissions and resource consumption associated to barley production can be divided in different general categories (i.e., acidification, eutrophication, climate change, and formation of photochemical oxidants) which are significantly affected by CO 2 , CH 4 , NH 3 , NO x , and N 2 O emissions. These aforementioned categories are the most commonly used in life cycle impact assessment (LCIA) studies, regardless of the type of agricultural system [37][38][39][40][41][42][43].

Sensitivity Analysis
In this type of estimation, the formulated assumptions are critical to the outcome and, hence, strongly influence the results and conclusions. Regarding the allocation method and the emission factors, several changes in the formulated assumptions were made to carry out a sensitivity analysis of the CC impact category. According to ISO 14044 [28], "whenever several alternative allocation procedures seem applicable, a sensitivity analysis shall be conducted to illustrate the consequences of the deviation from the selected approach".
Our baseline scenario was based on two assumptions: (i) we are dealing with a massbased allocation system where the system loads between the product mass (barley grain) and the by-product mass (barley straw) and (ii) the IPCC 2006 Guidelines [17]. Estimates were calculated for both wet and dry regions. The sensitivity analysis was carried out in order to compare the baseline with two alternative scenarios: (i) allocation based on economic criteria and (b) the IPCC 2019 Guidelines [16]. Table 4 summarises the baseline (BS1) and these two alternative scenarios (SC).

Results
Firstly, following the methodology depicted in Tables 2 and 3, the following emissions were estimated and included in the LCA. Nitrate (NO 3 − ) emissions due to leaching and runoff in wet regions were estimated as follows: 210 and 167 kg NO 3 − ha −1 per 2006 IPCC and 2019 IPCC, respectively, due to the Frac LEACH update that was carried out in the 2019 IPCC Guidelines. Regarding the NH 3 emissions in the air, the same emission factor was used for both IPCC Guidelines, with an estimated value of 6.53 kg ha −1 . With respect to NO emissions, the value used in the baseline (BS1) and alternative (SC1, SC2, and SC3) scenarios was estimated at 6.14 kg ha −1 . After applying the SALCA-P model, the phosphate emissions (PO 4 3− ) were estimated at 0.21 (P gw ) and 0.64 (P ro ) for phosphorous leaching to groundwater and phosphorous run-off to surface water, respectively. Finally, air, water, and soil emissions from the pesticide were included in the LCA: 4.80 × 10 −2 (air), 4.08 × 10 −2 (water), and 3.62 × 10 −1 (soil).

Environmental Impacts of Barley Production in Araba/Álava
The effect of the climatic area (wet and dry) is only shown for the climate change (CC) category, as this is the only one on which the different climate has an effect.
The field operations stage was one of the main contributors to all the environmental impacts; for instance, the field operations were responsible for 87% of OD, 77% of HTnoc, 91% of HTC, 85% of FEx, and 97% of MFRD (Figure 3). This stage is important in PM (almost 40%) and contributes around 20% to the CC impact category.
In this study, NH 3 emissions derived from mineral fertiliser application were the main contributors to TA (52.1%) and TE (51.8%), while NO 3 − emissions were the main source (87.3%) of ME. In terms of the mineral fertiliser type, the CAN and NPK mixtures contributed 22% and 30% for TA and TE, respectively. The FE impact category was mainly determined (70%) by the phosphate emissions, followed by agricultural field operations (19%). Only five impact categories were affected by NO x : 5% of PM, 55% of POF, 15% of TA, 19% of TE, and 6% of ME.
For the CC impact, the relative contribution of the main inputs required for barley production in Araba/Álava differed only slightly by climatic region (W and D). The production of fertilisers was the main contributor to the CC impact (about 40%) in both regions. Fertiliser production also played an important role in other impacts, such as TA (19%), TE (18%), and PM (15%). Direct and indirect N 2 O emissions accounted for almost 40% of the CC impact in wet regions and 35% in dry regions (indirect emissions factors were not applied in the dry regions). The results are discussed in more detail in the following section.
The pesticide production and the pesticide emissions were a minor contribution to all the environmental impact categories (<4%).

Relative Contribution of Agricultural Management Inputs and Activities on Climate Change Impact
As mentioned in the previous section, fertiliser production was the main contributor to the CC impact in barley production in Araba/Álava, both in dry and wet areas (38% and 41%, respectively), with CAN production being the largest contributor (25% on average) ( Figure 4). Secondly, direct N 2 O emissions from N application to soils constituted 30% of the total emissions. The largest contributors to N 2 O emissions were those from CAN fertilisation (17-19% of the total emission), followed by those from NPK fertilisation (9%), due to the different quantities of nitrogen per fertiliser type.
Regarding diesel combustion for barley crop production, the results have been disaggregated into three different groups, with field preparation and harvesting practices contributing a similar amount (8% on average), and cultivation practices contributing only 4%.
In this study, the indirect N 2 O emissions made up 9% and 3% of the contributions to the climate change impact for wet and dry regions, respectively.
Finally, pesticide and seed production were minor contributors (<4% for both wet and dry regions) to the climate change impact.

Sensitivity Analysis of the Climate Change Impact Category to Allocation Types and Emission Factors
The main objective of the sensitivity analysis was to investigate the effect of the choice of different allocations systems (mass-based vs. economic allocation) and IPCC methodologies (IPCC 2006 vs. IPCC 2019).
When economic allocation was considered instead of mass-based allocation in the wet regions of Araba/Álava (SC1-W), the CC impact worsened, due to the increase in the allocation factors obtained for grain, from 50% (by mass) to 80% (by economic). Hence, the CC impact increased by 80.6% ( Figure 5). In dry regions (SC1-D), there was a similar increase from 212 to 382 kg CO 2 eq t −1 (80.2%).
In general, the IPCC 2019 methodology significantly increased the CC impact in SC2-W (by 12%), SC3-W (by 103%), and SC3-D (by 48%), as compared to their corresponding baseline scenarios (BS1). These increases are the result of the new emission factors developed in the IPCC 2019 Refinement, which is based on the latest advances in science for direct and indirect N 2 O emissions. Only in the case of SC2-D was the CC impact reduced (18%), as compared to the BS1-D.
When we compare the baseline to scenario 2 (IPCC 2019 and mass allocation), the direct N 2 O emissions increased in wet regions and decreased below the baseline in dry regions ( Figure 6). The reason why the wet region's direct N 2 O emissions increased is because the EF 1 factor updated from 1% to 1.6% in wet regions when synthetic fertilisers were applied to the soil. In dry regions, the EF 1 was updated from 1% to 0.05% for all nitrogen inputs, leading to a deep decrease in the contribution of emissions. However, the emissions from the N in crop residues decreased by the decrease in EF from 1% to 0.06% kg N 2 O-N kg N applied −1 .  In the case of indirect emissions, the new emission factors decreased the relative contribution of the indirect N 2 O emissions to CC in both climatic regions. When comparing the baseline scenario to scenario 2 (mass allocation + 2019 IPCC), the emissions from leaching and runoff in wet regions decreased from 6% to 4%, while the N 2 O emissions from atmospheric deposition decreased from 3% to 1%. In dry regions, no leaching or runoff was calculated, but the atmospheric deposition decreased.
The other processes (seed, fertiliser, and pesticide production and diesel consumption) reduced their relative contributions to CC in wet regions due to the increase in direct N 2 O emissions. The opposite occurred in dry regions due to the decrease in both direct and indirect N 2 O emissions.

Environmental Impacts of Barley Production in Araba/Álava
In this study we found that the climate change impact was 220 kg CO 2 eq t −1 in barley produced on average for the baseline scenario. This result is lower than that obtained by Lovarelli et al. (2020) [10] (279 kg CO 2 eq t −1 ), who used the ILCD midpoint method [45] and performed a mass allocation to quantify the environmental impact of barley cultivated in Spain. Our results, when an economic allocation was applied (496 kg CO 2 eq t −1 on average), are higher than those obtained by Rivera et al. (2017) [12], who used the same ILCD methodology and implemented an economic-based allocation method for barley production in Italy and Denmark (265 and 345 kg CO 2 -eq t ha −1 , respectively). Our study is useful as it allows for the direct comparison of the different methodologies used in different studies, which used one methodology each.
The different methodological approaches between the present and other recent studies underline the difficulties involved in comparing LCA studies. Thus, a study carried out by Bartzas et al. (2015) [46], who used a different LCA methodology (CML [47]), determined that barley production in Spain has an overall impact of 171 kg CO 2 -eq t ha −1 , 0.67 kg SO 2 -eq for PA, and 0.56 PO 4 -eq for PE, with a grain yield of 5.42 t ha −1 .
Another difference between the present and other studies is mainly due to the amount and type of the fertilisers applied in the field. For instance, Lovarelly et al. (2020) [10] highlighted the improvement of the results on the impact categories affected by fertiliser production when animal slurry is applied, while worsening those affected by the N emissions. In the case of the use of mineral fertilisers, all the environmental effects related to ammonia volatilisation are reduced. This reveals that there is a trade-off between both scenarios (mineral vs. organic fertilisation) that must be considered when farmers, advisors, and policy makers are deciding upon a fertilisation regimen. However, the environmental impacts are largely affected by yield, which in the present study is quite high (6.2 t ha −1 ) in comparison to other studies (3.13 and 6.1 t ha −1 [10,12,46]).
Another factor that affects several environmental impacts is the level of machinery used and the definition of the boundaries and type of LCA that the study carried out. In the present study, diesel combustion from different field operations averaged 75.29 kg diesel ha −1 . Lovarelli et al. (2020) [10] set a value of 64.2 kg diesel ha −1 for the Spanish barley production, including straw windrowing and straw baling, in addition to associated transport, which were not included in the present study. The agricultural field operations stage was one of the main contributors to all of the environmental impacts, as found by Rivera et al. (2017) [12] and Lovarelli et al. (2020) [10].
In regard to field emissions related to barley cultivation, differences could be found between the emissions that are included in the studies. As Rivera et al. (2017) [12] summarised, most studies on cereals include N 2 O, NH 3 , and NO 3 emissions [11,13,43,[48][49][50], but many do not consider the contribution of NO x [4,51]. In our study, the emissions of NO x had an effect on five impact categories: particulate matter formation, photochemical ozone formation, terrestrial acidification, terrestrial eutrophication, and marine eutrophication. NOx was responsible for more than half the impact of POF. This demonstrates that not considering NO x emissions in the LCA inventories reduces their accuracy. This is in agreement with Roer et al. (2012) [40], who found that exclusion of NO x loss from the use of mineral fertiliser resulted in high reductions of PMF (65-71%), POF (83-82%), and TA (68-75%).

Relative Contributions of Inputs and Activities to Estimated Climate Change Impact
Our results are in accordance with previous studies, which identify field emissions and fertiliser applications as the main contributors to the estimated climate change impact [11,40]. Nitrogen fertiliser production is highly energy intensive; thus, fertiliser production in this and other studies was the greatest contributor to the barley production estimated climate change impact and accounted for an average of 40% of the estimated impact in this study. Furthermore, the direct N 2 O emissions implied an average of 31% of the contribution to the CC impact. Fertiliser production and N 2 O emissions are priority areas for focusing efforts to reduce greenhouse gas emissions from barley production. In this regard, other authors have focused on studying the improvement of the N uptake efficiency. These authors observed CC impact decrease due to a combination of higher yield, soil carbon sequestration, and lower indirect emissions of N 2 O due to lower N leaching [8]. According to Hasler et al. (2015) [52], with an optimised fertilisation strategy, the environmental burden can be reduced by up to 15%. Other previous studies showed that N 2 O emission can be significantly reduced by using non-nitrate fertilisers, because high amounts of N 2 O are emitted during the industrial production of nitric acid, which is part of the ammonium nitrate production [53,54]. High N use efficiency and lower N input might be key factors to minimise GHG emissions.
In addition, indirect N 2 O emissions contributed 9% and the other processes contributed less than 4% to the estimated climate change impact (seed production and pesticide production and emissions). The emissions of N and P compounds related to the mineral fertilisers are the main contributors to TA and TE (86% and 89%) and FE (70%), mainly due to ammonia volatilisation and phosphorus run-off, respectively, in accordance with the results obtained by other authors [10,12].

Sensitivity Analysis
Firstly, in comparison to the mass-based allocation method suggested by the product category rules (PCR) for arable crops, and consistent with other LCA studies [55,56], an economic-based allocation method has been evaluated; this is because, over a 5-10 year period, the price variability of the grain in Araba/Álava was reduced. When economic allocation was applied instead of mass-based allocation, all impact categories increased by 80% due to the higher load allocation to the main product (barley grain), as reported in other studies [4].
Secondly, the results of the sensitivity analysis underline how the IPCC 2019 Refinement method affects the climate change impact of barley production. The climate change impact increased by 1.12 times in wet regions, while it decreased 0.82 times in dry regions. However, the rest of the impact categories were not influenced as the methodological refinement only influences the calculation of direct and indirect N 2 O emissions. The IPCC 2006 default values assume a linear relationship between the amount of N applied to the soil as fertiliser or crop residue and the amount of N 2 O emissions [8]. However, the IPCC 2019 Guidelines identify the influence of the climate type, the N fertiliser type, and the influence of irrigation in dry regions.
Consequently, the emission factor disaggregation by climate and fertiliser (nitrogen) type resulted in an increase in estimated direct emissions of N 2 O in wet regions, while a significant decrease of estimated emissions was calculated in dry regions. The estimated indirect emissions of N 2 O decreased in both climatic regions.
Therefore, the results of the sensitivity analysis underline how the allocation method and the IPCC guideline selected significantly affects the estimated impact of CC of the barley production in Araba/Álava.

Future Research Directions
Even though the study of S. González-García et al. (2016) was focused on the production of animal feed, according to the results reported, a selection of alternative functional units could be proposed in future works since different units could be used in agricultural systems [40]. For example, (i) 1 ha; (ii) 1 t crude protein; and (iii) 1 MJ metabolisable energy. It would also be desirable to study the consequences of implementing the energy allocation system in the environmental assessment of barley production as an alternative or complementary method, based on the low heating value (LHV) of the product and byproduct (27 MJ kg −1 and 16 MJ kg −1 for the grain and straw, respectively). However, and as indicated in the PCR of arable crops [18], the allocation could be avoided by introducing a system expansion method. In that case, the avoided production of other bedding mate-rial for livestock use could be considered as a consequence of the straw production and, consequently, its environmental impact could be subtracted from the barley production.
The accuracy of the LCA results here could be improved by applying more robust models that require specific regional information in order to be applied, such as temperature, annual precipitation distribution, wind speed, month of fertiliser or pesticide application, soil characterisation (e.g., pH, texture, soil organic matter, etc.), pesticide type, crop type, stage of the development of the crop, and application method [12,34].
Further research could also be directed to include the environmental impacts assessment of the barley production in Araba/Álava under current and future climatic scenarios. In addition, it could help farmers and policy makers establish solutions to mitigate the environmental impact of crop production by adjusting the fertiliser application rate according to fertiliser type and climate region, and to establish adaptation measures through climate change [11,13].
Finally, conventional LCA only considers the emissions of the agricultural products, and not the incomes, such as the carbon sequestration in the products and the net increase in soil organic carbon (SOC). Hence, instead of conventional LCA, a carbon sequestration LCA accounting method should be considered to value the changes in SOC stocks, allowing for the estimate of net emissions instead of gross emissions. Several soil accounting LCA methods could be used, based on user expertise and data availability, as summarised by Goglio et al. (2015) [57].

Conclusions
The environmental impacts of barley production in Araba/Álava were evaluated using the LCA approach. For this purpose, the most representative cultivation practices were identified through farmer survey data. This study highlights how the choice of allocation system and IPCC methodology is a key issue when agriculture emissions are estimated. In this regard, the economic allocation increased all the impact categories due to the increase in the percentages attributed to the grain relative to the by-product (straw), as compared to the mass-based allocation method.
The 2019 IPCC Refinement Guidelines allowed for the delivery of more robust estimations of the climate change impact, as compared to the conventional IPCC 2006 methodology. Thus, relative estimates increased in wet regions and decreased in the dry regions by a factor of 1.12% and 0.86%, respectively. The disaggregation of emission factors in the new guidelines provides policy makers with reliable information to establish better fertiliser use guidelines to properly account for nitrogen emissions from barley production.
Fertiliser production, diesel combustion, and N 2 O direct emissions are the main contributors to GHG emissions. Regarding other environmental impacts analysed, the emissions of water-soluble N and P compounds related to mineral fertiliser applications are the main contributors to PM, TA, TE, and ME. Furthermore, the inclusion of NH 3 , NO x , NO 3 − , and PO 4 3− in the LCI makes it more accurate. Finally, the results obtained in the present study, in conjunction with other studies, can help policy makers establish mitigation options for the processes that affect the environmental sustainability of barley production in regions like Spain.  Data Availability Statement: The study did not report any additional data.

Acknowledgments:
The authors would like to acknowledge the work of the three anonymous reviewers in helping us to improve the article.

Conflicts of Interest:
The authors declare no conflict of interest.