The potential of behavioural change for climate change mitigation: a case study for the European Union

Abstract

Mainstream literature on climate change concentrates overwhelmingly on technological solutions for this global long-term problem, while a change towards climate-friendly behaviour could play a role in emission reduction and has received little attention. This paper focuses on the potential climate mitigation by behavioural change in the European Union (EU) covering many behavioural options in food, mobility and housing demand which do not require any personal up-front investment. We use the Global Change Assessment Model (GCAM), capturing both their direct and indirect implications in terms of greenhouse gas emissions. Our results indicate that modest to rigorous behavioural change could reduce per capita footprint emissions by 6 to 16%, out of which one fourth will take place outside the EU, predominantly by reducing land use change. The domestic emission savings would contribute to reduce the costs of achieving the internationally agreed climate goal of the EU by 13.5 to 30%. Moreover, many of these options would also yield co-benefits such as monetary savings, positive health impacts or animal wellbeing. These results imply the need for policymakers to focus on climate education and awareness programs more seriously and strategically, making use of the multiple co-benefits related with adopting pro-environmental behaviour. Apart from that, the relevance of behavioural change in climate change mitigation implies that policy-informing models on climate change should include behavioural change as a complement or partial alternative to technological change.

Appendices

Appendix 1: Background modelling of mobility and housing options

Public transport commuting

In the base year (2010), around 20.7 and 17.9% of total passenger kilometres in respectively EU-15 and EU-12 were due to commuting between home and work. Of this commuting transport demand, only 15.1% in EU-15 and 29.3% in EU-12 was met by public transport services (EC 2010a).

Carpool commuting

In the base year, car trips yielded around 76.2% of the total commuting passenger kilometres in EU-15, while around 58% in EU-12. Car load factors²⁹ for commuting transport were 1.19 in EU-15 and 1.87 in EU-12, whereas car load factors for total car transport were 1.65 and 2, respectively (EC 2010a). Assuming a car load factor for commuting transport of 2 while respecting the share of commuting kilometres in total passenger kilometres (20.7 and 17.9% in EU-15 and EU-12 respectively), we increased the overall load factor for all four-wheel driven transport modes to 1.85 in EU-15 and 2.05 in EU-12 to model this behavioural option. Finally, in order to model only the emission savings as a result of this behavioural change, we cancelled out any kind of price elastic behaviour in favour of car transport following to this adjustment.

Teleworking

Since the commuting share of passenger transport is 20.7 and 17.9% in the EU-15 and EU-12, respectively (see previous options), we reduced total passenger transport demand by 4.14% in EU-15 and 3.6% in EU-12.

Urban cycling

According to EC (2010a), slow mode transport (walking and cycling) accounts for 18.9% of total urban passenger transport in the Netherlands in 2010, while total urban passenger transport accounts for about 29% of all passenger transport. Assuming the same percentage of slow mode transport in urban areas for the whole of EU-27, this comes down to an average share of 5.4% of total passenger transport that would be met by walking and cycling together. GCAM reports the share of walking to account for 1.9% of passenger transport in EU-27 in 2010, so we assume that the potential share of bicycles in total EU-27 passenger demand will be around 3.5%. Note that while we keep the cycling share to 3.5% during all periods for this behavioural option, the walking share is subject to market competition (and decreases rapidly due to an increasing cost of travel time, see also, Mishra et al. 2013).

Car sharing/car clubs: methods and assumptions

The calculation used to make assumptions for both emission-saving effects is as following. Based on a ratio of 27 members per shared car in the USA, the Trasportation Research Board (2005) reports an amount of 14.9 cars to be taken off the road for every car-club vehicle. Applying the ratio of 20 members per shared car in Europe, the estimate for Europe would be 11 cars per car-club vehicle. Correcting this estimate by the 40% reduction in vehicle kilometres of car-share members compared with private vehicle owners, this ratio comes down to 11 × 0.6 = 6.62. Finally, Chen and Kockelman (2015) state that the average privately owned new vehicle is replaced after approximately 6 years, whereas commercial car-club operations replace cars every 2 to 3 years due to more vehicle kilometres and faster wear and tear (Mont 2004). Assuming that the wear and tear to the car and the remaining life time is the same for privately sold second hand cars and those sold by car-sharing companies, we can state that a privately owned vehicle has two to three times the lifetime of a car-club vehicle. Applying a lifetime ratio of 2.5,³⁰ this means that every car-sharing vehicle takes 6.62/2.5 = 2.65 vehicles off the production line when assuming that there is no reduction in car use between car owners and car sharers. Furthermore, we assume an energy consumption related to car manufacture of 30 GJ vehicle⁻¹ (Sullivan and Wang 2010) and a growing demand for cars proportionally to the growing demand for passenger kilometres in both EU-15 and EU-12. See Table 9 for a summary on the assumptions made for modelling the impacts of car sharing.

Table 9 Assumptions made to model car-sharing impact

Avoid short flights

We summed all the passenger kilometres on national flights within EU-27 member states³¹ and all flights to neighbouring countries (multiplied by half if at least one of the partner countries is a large country such as Germany, France, UK, Italy or Spain³²) to have a rough estimate of the potentially avoidable flights. We found that about 25% of all passenger kilometres on intra-EU flights are avoidable by these standards, and implicitly assume that it remains 25% until 2050.

As an alternative to flying for medium distance trips, we modelled a new category with four possible travel alternatives: coach, train, high-speed-rail and carpooling. Although we copied these transport modes from the original GCAM model, we assume significantly higher speeds for long distance bus, train and car transport (80, 100 and 100, respectively) and a higher load factor for cars.³³ Initially, each of these alternatives takes an equal share of the passenger kilometres to be replaced, but the mix between technologies is subject to mode competition as in other GCAM sectors.

Closer holidays

A rough analysis of Eurostat data on intercontinental passenger kilometres from EU-15 and EU-12 shows that respectively 85 and 91.5% of passenger kilometres are for leisure purposes and that the average intercontinental leisure trip by EU-15 and EU-12 consumers is respectively about 5900 and 2680 km long. We implicitly assume that these estimates will not change until 2050.

Eco-driving

According to the ecoDriver project website, the EU initiative that started in 2010 to promote this fuel-efficient driving style, the long-term fuel reduction due to eco-driving is estimated to be 5%.³⁴ Following this number, we modelled this behavioural option by increasing the efficiency of all four-wheel light duty vehicles by 5% from 2015 onwards.

Reduce heating and cooling

To model the reduced usage of heating, we simply modified the residential HDD input (heating degree days) from 4920 to 4625 in EU-15 and from 6311 to 5930 in EU-12, a change that reduces the need for heating in winter by about 1 °C. Similarly, we changed the residential cooling degree days (CDD) input from 373 to 328 in EU-15 and from 343 to 302 in EU-12 to model a reduced use of air-conditioning in summer.

Appendix 2: Modelling and assumptions of GCAM waste module

To model the impacts of waste recycling by consumers, we focus on the three main streams of consumer waste: organic waste, paper/carton waste and non-paper packaging waste (consisting of mainly plastics, metals and glass). In most EU member states, it is possible for households to effectively recycle these types of waste by separating them. For modelling simplicity, we will assume from now that 100% of separated waste actually will be recycled (8% actually ended up between mixed waste in 2010, predominantly separated organic waste in landfills) and that 0% of mixed waste will be recycled (8% of mixed waste was actually recycled in 2010). See Fig. 6 for an overview of all waste and recycling streams in EU-27 in 2010.

Since 66% of household waste ended up between mixed waste in 2010, it is hard to determine the contents of these waste streams. Since we need to know the contents to model the potential emission reductions, we have to make an estimation of these contents. To do so, we looked at the best practice example of waste separation in Europe to gain information about the average household waste streams. According to GAIA (2012), European best example is a door-to-door waste collection program in Usurbil, Hernani and Oiartzun in the province of Gipuzkoa, Basque Country, Spain. The three towns together represented 33,628 citizens with a GDP per capita level close to the EU-27 average. Except for the 20% of waste that was collected from street bins and local street cleaning services, all household waste in these villages was separately collected. The household waste in these villages consisted of 46.8% organic waste (of which 33.8% food and 13% garden waste), 18.3% paper/carton waste, 32.3% industrial packaging waste (including 14.1% glass and 15.2% plastic and metal) and 2.6% other waste, such as chemicals or minerals

Since all EU-27 member states have a different waste collection scheme with regionally different priorities, we have multiplied the household waste composition as assumed above to the waste totals in every member state and have deducted the separated waste streams from these assumed waste streams. The remaining waste (i.e. the composition after deducting the separated waste streams per member state) is assumed to be the composition of waste within the mixed waste stream. On an EU-27 level, we find 45.6% of all mixed household waste to be organic, 13.6% to be paper/carton, 33% to be non-paper packaging waste and a remainder of 7.8% to be mineral or chemical waste (which we leave out of the model).

Fig. 6

EU-27 waste and recycling streams in 2010 in million tonnes (based on EuroStat data)

For the services and industrial sector (accounting for nearly one third of all mixed waste), waste has traditionally been much better separated. We therefore assume the same mixture of separated waste to hold for the limited amount of mixed waste streams from these sectors. Finally, we also find that about one fifth of the mixed waste in the waste collection industry. This is intentionally separated waste that has a degree of mixture too high to be recycled. Here we simply assume the average assumed waste composition as in the other 80% of mixed waste. The final assumed mixed waste contents in EU-27 are assumed to be 34.3% organic waste, 15.4% paper/carton waste, 31.2% non-paper packaging waste and 19.1% other waste (mainly mineral). Note that we only modelled the non-household sectors to have a full picture on all waste streams. All the behavioural options do apply to household waste only.

In total, 89% of all mixed waste in EU-27 was treated within the area,³⁵ with the majority being landfilled (in some cases with methane recovery for biogas production). However, there is an important trend going on in Germany, the Benelux and Scandinavia to incinerate mixed waste, either with or without energy recovery. In the baseline estimates, we assume that open burning and unmanaged landfilling of waste will be phased out linearly until 2050 and also managed landfilling will be phased out linearly until 2100, following Directive 2008/98/EC on waste management.

For the total emissions from landfilling, we used data from the European Environment Agency (EEA)³⁶ on landfill emissions on managed and unmanaged landfill sites. Following the IPCC guidelines (IPCC 2006), unmanaged landfill sites have on average 40% less emissions per unit of waste compared with managed landfills.³⁷ For modelling simplicity, we assume that all landfill emissions in one period are coming from waste that is landfilled in the same period. To fit our modelled waste streams (stemming from EuroStat data) with the EEA landfill emissions data, we use the methane yields per type of waste stream from EPA (2015). Following the IPCC guidelines, we do not model CO₂ emissions from municipal waste management.

Organic waste

Organic waste consists of both food waste and garden waste. Since we have modelled the assumed amount of food waste in EU-27, we assume that all other organic household waste consists of garden waste (the relative share of garden waste is in line with the distribution in our case example as explained in the start of this Appendix). From 2010 onwards, we assume per capita garden waste to remain constant over time. Food waste consists of unavoidable food waste (which is a by-product of food consumption, predominantly skins and peels of fruits and vegetables, carcasses of pork and chicken, coffee and tea disposals) and avoidable food waste from the production, distribution and consumption of food. We estimated the unavoidable waste stream by GCAM food category by connecting the share of unavoidable waste compared with avoidable waste as reported by Ventour (2008) with our FAO’s’s (2011) estimates of avoidable food waste by food category. Estimates for unavoidable coffee and tea waste streams (NonFood-MiscCrop) come from van Westerhoven (2013a). See the assumed estimates in Table 10.

Table 10 Assumed unavoidable waste streams from different food categories (% of total weight)

Landfilling of organic waste results in large amounts of methane due to the anaerobic decomposition of organic materials. These are responsible for 2.75% of total GHG emissions in EU-27 in 2010. When incinerated, there will be no methane emissions from organic waste but there will be CO₂ emissions, which have a significantly lower warming potential.³⁸ The energy density of organic waste, however, is very low, so energy recovery from incineration is not very productive. Finally, the preferred treatment for organic waste is to compost it using anaerobic digestion, creating both biogas and a valuable organic fertiliser replacing mineral fertilisers and returning about 15% of the organic carbon contents back into the soil. This is a form of carbon sequestration (IPCC 2006). Some methane emissions are released in the composting process, but these are limited compared with the methane released with landfilling. In the same way as landfill emissions, we linked data from EEA on composting emissions with Eurostat data on total tonnes composted to estimate the methane and nitrous oxide emissions per unit of food and garden waste composted. Finally, we used estimates from Boldrin et al. (2010) and Zero Waste Europe (2015) to estimate the total carbon and nitrogen content of both food and garden waste.

Paper/carton waste

We have separated paper waste, since nearly every EU member state offers the possibility to recycle paper and carton waste. Since paper products are made from pulp, which is obtained from forest products, the GCAM model will be helpful in calculating the emissions related to paper waste recycling. Like food and garden waste, paper waste is organic and therefore leads to methane emissions when landfilled. However, the rate in which one ton of paper waste produces methane is only about one fourth compared with that of food waste (EPA 2015). When incinerated, paper products can yield significant energy recovery due to an energy density that is more than twice that of food and garden waste. Finally, recycling of paper waste leads to significant GHG savings: producing new paper out of recycled paper reduces the amount of energy needed for paper production by 40% (EIA 2006). However, since about four fifth of this saved energy comes from biomass (black liquor) due to the high amount of wood waste in these production locations (Table 17 in AF&PA 2009), paper production from pulp consumes the majority of the biomass energy in the EU-27 energy mix for industrial products.

Plastic/metal/glass waste

Although industrial products such as plastic, metal and glass do not emit GHG emissions when landfilled, they do emit other pollutants, which are currently not modelled within GCAM. These pollutants are also emitted when incinerated, along with CO₂. Glass and metal waste might also lead to health damages or complicate the whole waste collection procedure by cutting into garbage bags due to their sharp edges. Incineration with energy recovery from predominantly plastic waste is interesting due to its high energy density: around 50% higher than paper waste and four times higher than food and garden waste. Plastic, metal and glass waste however is most valuable when recycled: compared with producing new products, using recycled plastic, metal or glass reduces industrial energy use by 70, 60–95 and 5–30%, respectively (the Economist 2007). Given the average mixed waste composition in the EU-27, we have estimated that the average tonne of recycled industrial products saves about 30% of industrial energy compared with making the same final industrial products from virgin material (Zero Waste Europe 2015). It is important to note is that the majority of savings comes from recycling metal waste, which saves 60 to 95% (for aluminium) compared with making these products from virgin materials.

Appendix 3: Sensivity analysis based on starting year of behavioural change

Table 11 Avoided GHG emissions per behavioural option or profile dependent on starting year of behavioural change