Calculation of the CO2 footprint using the example of Austrian wine

The current standards for voluntary reporting of GHG emissions (GHG Protocol, ISO 14064-1) leave a great deal of freedom in the selection of calculation methods and data sources. Therefore, system boundaries must be clearly defined and documented for each study. According to ISO 14067 standards, the life cycle stages to be investigated in the balance are defined by the following system boundaries: i) Cradle-to-Grave: Includes emissions and removals that occur throughout the life of the product. ii) Cradle-to-Gate: Includes emissions and distances to the point where the product leaves the organization. In this work, the GHG emissions were analysed within the framework of “cradle-to-gate” of important production steps over the entire production chain of the product (0.75l bottle).

GEMIS is a freely available computer model for life cycle and material flow analyses in analyses for ecological damage. It was developed by Öko-Institut e.V. (Institute for Applied Ecology, Freiburg, Germany) and was created with funding from the Hessian Ministry of Environment and Economics in its first version in 1989. Since then, it has been continuously updated and expanded with funding from, among others, the German Federal Ministry for the Environment, the German Federal Ministry of Research, as well as the German Federal Environmental Agency, the GIZ, the EEA, and EU projects. In April 2012, GEMIS was transferred to the International Institute for Sustainability Analysis and Strategies (IINAS), which will take over further development and data maintenance [GEMIS, 2021].

Based on existing research work, the Austrian Umweltbundesamt GmbH has further developed GEMIS with the aim of generating greenhouse gas and air pollutant balances for Austria with country-specific adaptation for energy and material processes. The adapted GEMIS tool takes into account all essential processes, starting from primary energy and raw material extraction up to useful energy and material supply, e.g. also auxiliary energy and material input for the production of energy plants and transport systems. It thus offers the possibility of considering not only direct emissions but also upstream process emissions, the so-called indirect emissions (GEMIS, 2021). The emission factors used for greenhouse gas balances in this adapted model are regularly compared to the data material from the Austrian Air Pollutant Inventory (OLI) and reflect the country-specific reality. Austria is obliged to compile an annual greenhouse gas inventory of all economic sectors (National Inventory Report – NIR, 2021). All calculations illustrated in the current study have been performed using this model.

The system boundary for the calculation of greenhouse gas emissions from a vineyard in Austria was set to the functional unit of one hectare of vineyard area in Austria and a wine yield of 6750 litres after fermentation on average. Vineyard area and the harvest volume are the basis for the energy and material inputs used. All energy and material inputs refer to this functional unit and one business year. The treatments between fermentation and bottling cause a loss of 7 %, which is taken into account in the calculations. This loss is generally assumed by Austrian tax authorities, and thus this quantity was adopted unreflectively from a technical point of view. The system boundary does not include business travel (air travel, rail travel), wine logistics (neither the company’s own nor third-party fleet or delivery companies (e.g. DPD), refrigerant losses from the refrigeration machines, employee travel, and infrastructure construction materials (wine cellars, buildings, warehouses, wine tanks, wine presses).

The current study is based on key performance indicators (KPI), namely relevant key figures allowing to define targets and plan suitable measures. Here, the KPI of a winery in Austria is defined as the total of GHG emissions in relation to the yield per year.

As illustrated above, an average grape yield with 9,000 kg and wine yield of 6750 litres minus 7 % treatment losses per ha of vineyard are taken as basis for the current calculations. Based on empirical values and, where available, data collections, the energy and material inputs for one ha of an average vineyard were compiled as illustrated in Table 1 and illustrated in detail below. These figures served as basis for the calculation of the KPI GHG emissions.

The infrastructure of an average Austrian vineyards includes 3570 stakes (0.1 kg/stick), 715 poles (5 kg/pole), and 6000 linear meters of wire (25 kg per 1000 m). The construction of one vineyard is related to the assumed vineyard life of 30 years. For the management and harvest of 1 hectare of vineyards, 160 litres of diesel per year are required (source: data analysis in poelz et al, 2020; Rosner et al., 2015). Plant protection is taken into account with 11 kg of conventional/systemic plant protection products. In the study at hand, seven applications with 1.5 kg of product per treatment were assumed. The use of nitrogen through mineral fertilizers entails two different greenhouse gas effects. On the one hand, the Haber-Bosch process allows a synthetic production of nitrogen fertilizers. However, this process requires a high energy input, which results in corresponding greenhouse gas emissions. On the other hand, nitrous oxide is emitted during nitrogen applications. This gas has up to 300 times higher greenhouse gas potential than CO₂. Nitrous oxide emissions are included in the calculations based on the amount of nitrogen added, regardless of the type of fertilizer (mineral fertilizer, farm manure, compost, or green manure). In a guideline issued by the Austrian Ministry for Agriculture, Forestry, Regions and Water Management, fertilizer input per hectare amounts to 40 kg of pure nitrogen, 70 kg of potassium, 20 kg of phosphorus, and 25 kg of magnesium per year (BMLFUW, 2020). These quantities have been included in the current assessment. In the winery, 175 kg of sugar are used to enrich the 6,750 litres by 2 degrees “Klosterneuburger Mostwaage” (KMW). Since the quantity refers to the fermented wine and not to the must, the increase in volume due to the addition of sucrose is not taken into account. For all must and wine treatments (protein, tannin and clarification), 19.5 kg of medium is factored in to the actual calculation. The by far largest part of this quantity is made up by bentonite. The energy input in the form of electricity in the wine cellar, during wine storage through to bottling, is around 2,400 kWh for 6,750 litres minus 7 % treatment loss in the winery. The electricity use of 2,400 kWh in the cellar up to bottling includes the following operations: grape pressing, must and wine pumping, cleaning work with steam cleaning equipment (e.g. Kärcher high-pressure cleaners); stirring of wine (addition of fining agent); sterilisation of the bottling plant. All energy quantities were defined as electricity. 2,400 kWh electricity cause 641 kg CO₂ equivalent emissions. This corresponds to an average electricity park in Austria. For the bottling of the wine produced in one ha of vineyard, 8,370 0.75l Bordeaux-style/claret bottles with a net weight of 480 g are assumed in the calculations. Packaging (488.25 kg for the cardbard wine cases at a weight of 0.350 g/6 cases), 15.7 kg labels and 25.1 kg aluminum capsule closures (4 kg per 1000 litres) are included in the system limit.

Our previous study revealed that the glass bottle accounts for 47 % of the CO₂ footprint generated during production (Poelz et al., 2020). In consequence, the current study aimed to investigate the impact of different glass bottles. The use of lightweight glass bottles as well as the use of glass bottles exclusively produced by fossil fuel energy were analysed. In addition, we investigated the potential of alternative packages for emission reduction. The packages outlined in Table 2 were included in the study.

Refilling of small containers requires a precise cleaning procedure before the packaging can be reused. In order to include the actual effect of bottle cleaning into our assessments we requested data for large-scale and modern machinery from the company Krones in Neutraubling, Germany (Krones, 2023), selling bottling and packaging equipment. Data per bottle during bottle washing in small-scale and standard operations were compiled based on empirical values. Case studies from Krones, Germany, were used to calculate the energy and material consumption on the one hand for bottle washing as part of a rinsing centre (appendix 1) and comparatively for the cleaning and filling of returnable glass bottles (appendix 2). As a basis for the calculation a refill rate of 12 times was assumed. It will also be investigated what GHG reduction occurs in a winery if the share of refilled bottles is 50 %.

The use of biofuels in Austria is controlled and must lead to a reduction of GHG emissions. The 2012 Fuel Ordinance (KVO) transposes the Directive on the Promotion of Renewable Energy Sources (28/2009/EC) into national law. The KVO provides for the use of biofuels iwith shares of 6.3 % for diesel fuel and 3.4 % for petrol fuel, respectively.(KVO, 2012). The use of biofuels must lead to a greenhouse gas reduction of 6 % in fuel use in Austria along the entire value chain. The Federal Environment Agency has set up the monitoring system elNa (electronic sustainability certificate), with which the mass balance can be complied with and checked. In Austria, biodiesel from unknown sources is not allowed to be traded. Due to these two facts, certified biodiesel is used in the GHG calculations.

Viticulture without fertilization is not successful in the long run. In consequence, renouncing fertilization was not considered as an alternative scenario in this study. However, synthetic fertilizers can be replaced by farm manure and the use of compost replaces the application of mineral fertiliser. Nitrogen fertiliser production in particular is very energy-consuming and thus emissions-intensive. The effect of this strategy – using compost instead of commercial fertiliser – was analysed in the study.

Based on the assumptions and data stated above, the GHG emissions of a winegrowing operation for 1 hectare in Austria amount to 6,591 kg CO₂ equivalent. The largest share of the GHG emissions is related to Scope 3, namely 82.1 %, 9.8 % are related to Scope 1 and 8.1% to Scope 2 (Table 3). Assuming a yield per hectare of 6,277 litres (6,750 litres minus 7 % loss), the KPI of a winery is 0.91 kg CO₂ equivalent emissions per litre of bottled wine or 0.68 per 0.75l bottle. Grape production accounts for 1,733 kg CO₂ equivalent per ha. Assuming a yield of 9,000 kg per ha, 1 kg of grapes would cause 0.19 kg CO₂e.

The biggest share of the emissions, namely 3,119 kg (48 %) are caused by the bottle packs. Fertilization (including the nitrous gas emissions) accounts for 667 kg, (10.3 %). Electrical energy use contributes 641 kg (9.9 %) and diesel use 521 kg (8.0 %) to the GHG emissions.Referring to a 30-year life cycle of a vineyard, annual material inputs for the establishment of a new vineyard amount to 417 kg (6.4 %). The closures including labels and cardboard boxes cause 710 kg (10.9 %). The enrichment with 263 kg (4.0 %), the plant protection products with 128 kg (2.0 %) and the wine treatment products with 30 kg (0.5 %) together make up 6.5 % of the GHG emissions. A calculation of GHG emissions as outlined above allows, based on the presented key performance indicators (KPI), an unequivocal identification of areas with high impact on GHG balances. In consequence, suitable measures can be planned and implemented. As an example, the current study highlights the enormous contribution of glass bottles to the total GHG emissions. The material input for 8,370 bottles is massive and can betracxed back to high emissions during production and transport of the bottles. Strategies to reduce this material use could greatly influence the GHG balance.

As already outlined, an average wine yield of 6,277 litres per ha of vineyard were taken as basis for the current calculations. The effective yield of a winery, however, massively depends on the vintage-specific conditions. The climatic conditions of a given vegetation period have a great impact on vine development, e.g. on flowering and bunch and berry size, in consequence, annual yields may vary greatly. Yield differences have a relevant impact on the GHG emissions per bottle or litre of wine, because input in many areas, such as vineyard management or harvest of the grapes occurs independently of harvest size. According to Statistics Austria, the 5-year average harvest in Lower Austria is 5380 litres per hectare (Statistik Austria, 2022). Compared to the assumptions for the model vineyard outlined above, the lower harvest can be filled in 7174 instead of 8,370 bottles and the required amount of sugar decreases from 175 kg to 140 kg. Based on the 5-year yield average GHG emissions amount to 5,771 kg per hectare and 1.04 kg of CO₂-equivalent emissions per litre of wine. One kg of grapes would therefore cause 0.24 kg of CO₂e.

The use of lightweight glass (370 g) instead of normal glass (480 g) leads to a reduction in material input without changing the filling capacity. The lightweight glass bottle lowers the material input by around 23 % and thus the GHG emissions to the same extent. All in all, the use of lightweight glass instead of normal glass diminishes the total GHG emissions of the model winery to 5630 kg per hectare (-12 % compared to normal glass in the model winery described above). Compared to other countries the rate of glass recycling in Austria (75 %) is outstanding. Assuming the use only of primary glass and exclusively fossil energy sources for the production of wine bottles, the emission factor for glass production would increase by about 48 %. This is based on the fact that 3% energy and 7 % CO emissions are saved for every 10% of used glass in new bottle production (vetropack, 2023). Under the assumptions of fossil-produced wine bottles, GHG emissions would come to 8012 kg (+25 % compared to normal glass in the wine operation described above).

The outcome of our model calculations, investigating the potential of alternative packaging for emission reduction, are outlined in Table 4.

The bag-in-box system can only be used once. This system leads to 59 % higher GHG emissions per litre compared to reusable bottles (system “Sustainability”). The higher material input for the “Exclusive” glass bottle (600 g, it requires 71 % more material than the “Sustainability” system (370 g)) is also reflected in the GHG emissions. The “Exclusive” system would therefore have to be used 7 times more often to achieve the same GHG emissions as the “Sustainability” system.

At a refill rate of 50 %, as shown in table 4 with a 370g/0.75l bottle and a refill rate of 12, GHG emissions are reduced to 4,367 kg per ha (-32 % compared to common 480g/0.75 l glass in the winery described above). The refilling of wine bottles represents the largest GHG savings effect in a winery. It can be unequivocally concluded that the refilling of small container systems is the essential step towards a wine industry with low greenhouse gas emissions.

As outlined above, the reuse of glass bottles is of crucial importance in a development towards low GHG emissions in wine production. However, reuse results in GHG emission due to bottle cleaning, which needs to be considered in a total balance. Calculations based on the data by the company Krones are illustrated in Table 4 and attachments 1 and 2. The GHG emissions for washing amount to 0.011 kg CO₂ equivalent emissions per bottle (emissions data from company Krones; Krones, 2022). Small or standard scale bottle washing are a little less environmentally friendly, in this case a value of 0.028 kg per bottle was calculated (Table 5). In any case, GHG emissions for bottle washing are far lower than emissions for the production of new bottles. For example, the production of one light glass bottle (370 g, Vetropack) emits 0.296 kg CO₂ equivalents, compared to the production of one standard bottle (0.75 l, 480 g, Vetropack) 0.328 kg CO₂ equivalent. All in all, the data indicate that despite the necessity for washing, the reuse of glass bottles remains by far the most effective measure in respect to GHG savings.

The Austrian recommended “Good Agricultural Practice” recommends mineral fertilizer (40 kg nitrogen, 70 potassium, 20 kg phosphorus, 25 kg magnesium) per year for viticulture. The use of farm manure in combination with permanent green cover containing legumes instead of mineral fertilizer can reduce GHG emissions to 6,045 kg per hectare (−5.5 % compared to mineral fertilization in the vineyard described above). However, a complete renunciation of fertilization is not a sustainable strategy in viticulture. Well-supplied soils are essential for long-lasting, vital vineyards and the production of ripe, healthy grapes. Farm fertilizers can supply both nitrogen and the main nutrients to the soil. Likely, a (partial) replacement of synthetic fertilizers by green or farm manure could significantly contribute to a GHG reduction.

The complete substitution of fossil diesel (160-l/ha) with biodiesel can reduce GHG emissions to 6,154 kg (−3.8 % compared to fossil diesel in the winery described above). An important point of discussion is whether the production of biodiesel has caused any direct land use changes (DLUCs). If the feedstock comes from agricultural land, indirect land use changes (ILUCs; Baral and Malins, 2016) must be taken into account. In this case, the cultivation leads to a shortfall in the harvest of other agricultural products that would otherwise have been cultivated on this land. This shortfall must now be compensated. This can be done either by ploughing up a new, previously uncultivated natural area for cultivation elsewhere. In this case, extensive biodiversity effects are clearly indirectly triggered. However, it is also possible that currently unused fallow land is used for this cultivation. Since fallow land in the agricultural landscape plays a crucial role in maintaining open land biodiversity, this variant is also clearly negative in terms of effects. The third possibility is that the lost area for biodiesel production is compensated for by a massive intensification of cultivation on existing land. Such intensification would be land-neutral, but it would require possibly the increased use of plant protection agents, which in turn would have negative effects. The problem of indirect land use change has been discussed extensively in the literature for some time (e.g. Kim and Dale, 2011; O'Hare et al., 2011). In addition to the sustainability effects, it must also be checked whether all machines are suitable for the use of biodiesel, which entails new investments for older types.