Exploring carbon footprint reduction pathways through urban lifestyle changes: a practical approach applied to Japanese cities

In this study, the household carbon footprint of multiple cities is estimated using an input–output approach together with mixed-unit consumption data. In many countries, monetary GHG intensity data from a single-country or multi-regional input–output model are available; however, for a more accurate modeling of lifestyle changes, reflection of physical consumption are needed. Insofar as household consumption and price data are widely available from expenditure and retail price surveys at the subnational level in many countries, this study combines country-level GHG intensity data and city-level price and consumption data.

Understanding the differences in physical demand among cities is useful for a more accurate modeling of demand-side changes such as the substitution of consumption items. In this study, the physical amount of household consumption in cities was estimated from city-level expenditure data and price information for items for which price information was available (e.g. mobility distances and energy at home) to reflect different local commodity price (e.g. gasoline and electricity prices). For items for which price information was not available, such as consumer goods, leisure, and services, monetary expenditure data were used. The units of the household consumption and GHG intensity data were hybridized using price data to estimate city-specific carbon footprints as follows:

$$\begin{align} {\text{C}}{{\text{F}}_j} & = \sum\limits_i \left( {{\text{H}}{{\text{I}}_i} \times {\text{H}}{{\text{C}}_{i,j}}} \right)\nonumber\\ & = \sum\limits_i \left( {{\text{M}}{{\text{I}}_i} \times P_i^{{\text{country}}} \times \frac{{{E_{i,j}} \times {\text{ca}}{{\text{j}}_i}}}{{{P_{i,j}} \times {\text{pa}}{{\text{j}}_i}}}} \right).\end{align} \tag{ 1 }$$

Here, ${\text{C}}{{\text{F}}_j}$ is the per-capita carbon footprint of household consumption in city j (in kgCO₂e/cap⋅yr), which is given by ${\text{H}}{{\text{I}}_i}$, the mixed-unit GHG intensity of item i (in kgCO₂e/physical-unit (e.g. kg, km-passenger, kWh) for selected domains such as food, mobility, and housing and kgCO₂e/monetary-unit for other domains (e.g. JPY)), and ${\text{H}}{{\text{C}}_{i,j}}$, the per-capita mixed-unit household consumption of item i in city j (in physical- or monetary-unit/cap⋅yr). The mixed-unit GHG intensity is given by the monetary intensity ${\text{M}}{{\text{I}}_i}$ (in kgCO₂e/monetary-unit) and country-level price data $P_i^{{\text{country}}}$ (in monetary-unit/physical-unit). The city-specific mixed-unit consumption ${\text{H}}{{\text{C}}_{i,j}}$ is given by the per-capita household expenditure ${E_{i,j}}$ (in monetary-unit/cap⋅yr) and price data ${P_{i,j}}$ (in monetary-unit/physical-unit), with adjustment factors for consumption amounts ${\text{ca}}{{\text{j}}_i}$ and prices ${\text{pa}}{{\text{j}}_i}$. Adjustments were made in the expenditures and the physical amount of consumption to match national level top-down data in order to address the possible underreporting bias in the bottom-up household survey data. As the purpose of this study is to explore footprint reduction pathways in comparison to the target based on an understanding of the carbon footprint hotspots and physical consumption patterns, such an adjustment is needed to reflect more realistically the carbon footprint levels and physical consumption patterns consistent with the national statistics. The total household carbon footprints are adjusted to the national inventory plus the emissions from international bunkers at the country level to meet consistency requirements, formulated as follows:

$$\begin{align}{\text{CF}}_{\text{c}}^{{\text{country top-down}}} = \sum\limits_{i \in C} {\left( {{\text{M}}{{\text{I}}_i} \times E_i^{{\text{country}}} \times {\text{ca}}{{\text{j}}_i}} \right)_{ }}.\end{align} \tag{ 2 }$$

The adjustment factor for consumption amount ${\text{ca}}{{\text{j}}_i}$ (in monetary-unit/monetary-unit) is determined for each component c (group of consumption items) to equalize the country-level carbon footprints estimated from the top-down input–output data ${\text{CF}}_{\text{c}}^{{\text{country top-down}}}$ (in kgCO₂e/cap⋅yr) to those estimated from the bottom-up household survey data $E_i^{{\text{country}}}$ by summing for all items belonging to the group of items under corresponding component C. The total household physical consumption amount is adjusted to the national statistics at the country level as follows:

$$\begin{align}{\text{HC}}_i^{{\text{country}}} = \frac{{E_i^{{\text{country}}} \times {\text{ca}}{{\text{j}}_i}}}{{P_i^{{\text{country}}} \times {\text{pa}}{{\text{j}}_i}}}.\end{align} \tag{ 3 }$$

The adjustment factor of prices ${\text{ pa}}{{\text{j}}_i}$ (in price/price) is determined to equalize the country-level household physical consumption amount based on national statistics ${\text{HC}}_i^{{\text{country}}}$ (in physical-unit/cap⋅yr) to that estimated from the bottom-up expenditure survey data $E_i^{{\text{country}}}$ and price data $P_i^{{\text{country}}}$ for each item.

With this methodology, city-specific household carbon footprints incorporating physical consumption can be estimated for multiple cities. In the case study, 52 Japanese cities (51 municipalities and the Tokyo metropolitan area) were selected to represent the variety of urban contexts in Japan by including all prefectural capitals and government-designated cities. Sources of the expenditure and price data used in the case study include the 2015 family income and expenditure survey (Statistics Bureau of Japan 2015a), which provides annual expenditure per household over 500 consumption items in JPY/household for each target city, and the 2015 retail price survey (Statistics Bureau of Japan 2015b), which provides the average price for a variety of items in JPY/physical-units in cities. (A comprehensive list of data sources is provided in tables S1 and S2 (available online at stacks.iop.org/ERL/16/084001/mmedia in supplementary information 1) (hereinafter, SI 1)). Using the price information, mixed-unit household consumption data in the target cities was estimated by incorporating physical units. This provides city-level consumption data for energy, housing space, mobility distance, and food (e.g. in kWh, m², km-passenger, kg), while consumption for other items (e.g. consumer goods and services) are kept in monetary units (JPY).

The monetary GHG intensity was obtained in kgCO₂e/JPY from the 2015 embodied energy and emission intensity data for Japan using input–output tables (3EID) (Nansai 2019, Nansai et al 2020), which is converted to mixed-unit intensity in kgCO₂e/physical-unit for food, mobility, and housing, and kgCO₂e/JPY for other items. The boundary of carbon footprint estimation in this study is the direct emissions of household activities and the indirect emissions through supply chains induced by the final demand of residents of the target city; the indirect emissions from imported products are estimated based on the domestic technology assumption. The adjustments in expenditure and physical consumption were made using government and industrial data sources, including the Japanese input–output tables (JIO) (Ministry of Internal Affairs and Communications 2015), the annual report on energy (Agency for Natural Resources and Energy 2015), and vehicle transport statistics (Ministry of Land Infrastructure Transport and Tourism 2015) (Details and data sources are included in tables S2 and S3). The per-capita carbon footprints of household consumption in the selected 52 cities as of 2015 were estimated. The estimated carbon footprints were classified into 42 components in six domains—mobility, housing, food, goods, leisure, and services—as specified in table S4. (A comprehensive data sources and methodologies is included in section 1 in SI 1.)

The carbon footprint reduction impact of lifestyle changes can be determined by the reduction or shift in consumption and changes in GHG intensity. The city-specific carbon footprint data incorporating the physical units of consumption estimated in the previous step are useful for a more accurate modeling of the shift of consumption from one item to another. To compare lifestyle change options and cities, it is important to first model the maximum impacts of the uptake of a particular lifestyle change option in a city, and then estimate the partial adoption of the same option and the combination of multiple options. Accordingly, the present study quantifies the carbon footprint reduction impact associated with a particular lifestyle change option as follows:

$$\begin{align}{\text{IMP}}_{j,k}^s = {\text{ CF}}_{j,k}^s - {\text{CF}}_j^b.\end{align} \tag{ 4 }$$

Here, $IMP_{j,k}^s$ is the per-capita carbon footprint reduction impact associated with lifestyle change option k in city j, which is given by the comparison of the scenario ${\text{CF}}_{j,k}^s$ and baseline carbon footprints ${\text{CF}}_j^b$, given as

$$\begin{align}{\text{CF}}_j^b = { }\sum\limits_i \left( {{\text{HI}}_i^b \times {\text{HC}}_{i,j}^b} \right)\end{align} \tag{ 5 }$$

$$\begin{align}{\text{CF}}_{j,k}^s = { }\sum\limits_i \left( {{\text{HI}}_{i,j,k}^s \times {\text{HC}}_{i,j,k}^s} \right).\end{align} \tag{ 6 }$$

The reduced per-capita carbon footprints ${\text{CF}}_{j,k}^s$ under the scenario assuming the adoption of lifestyle change option k in city j are given by the scenario GHG intensities ${\text{HI}}_{i,j,k}^s{ }$ and consumption amounts ${\text{HC}}_{i,j,k}^s$ as follows:

$$\begin{align}{\text{HI}}_{i,j,k}^s = {\text{ HI}}_i^b \times \left( {1 - {\text{ S}}{{\text{I}}_{i,k}} \times {R_{j,k}}} \right){ }\end{align} \tag{ 7 }$$

$$\begin{align}{\text{ HC}}_{i,j,k}^s = {\text{ HC}}_{i,j}^b \times \left( {1 - {\text{ S}}{{\text{C}}_{i,k}} \times {R_{j,k}}} \right) + {\text{HC}}_{i,j,k}^{{\text{sub}}}\end{align} \tag{ 8 }$$

$$\begin{align}S{I_{i,k}},{\text{ }}S{C_{i,k}} \leqslant 1{\text{ }}\end{align} \tag{ 9 }$$

$$\begin{align}0 \leqslant {\text{ }}{R_{j,k}} \leqslant 1{\text{ }}\end{align} \tag{ 10 }$$

Scenario GHG intensities ${\text{HI}}_{i,j,k}^s$ and consumption amounts ${\text{HC}}_{i,j,k}^s$ are given by intensity reduction factors ${\text{S}}{{\text{I}}_{i,k}}$, amount reduction factors ${\text{S}}{{\text{C}}_{i,k}}$, and the adoption rates ${R_{j,k}}$ applied to the baseline GHG intensities ${\text{HI}}_i^b$ and consumption amounts ${\text{HC}}_{i,j}^b$, with consideration given to the increased consumption of some items for substitution ${\text{HC}}_{i,j,k}^{{\text{sub}}}$. To illustrate, the maximum impacts of energy saving from nudging are modeled as reducing the energy consumption for heating and cooling in kWh by amount reduction factors ${\text{S}}{{\text{C}}_{i,k}}$, identified from the existing literature. Similarly, the impacts from the maximum adoption of electric vehicles are modeled as reducing the GHG intensity of driving by intensity reduction factor ${\text{S}}{{\text{I}}_{i,k}}$, identified from the improved fuel efficiency compared to the existing composition of vehicles. Reduction factors ${\text{S}}{{\text{I}}_{i,k}}$ and ${\text{S}}{{\text{C}}_{i,k}}$, with values up to 1, are determined based on the theoretical maximum uptake of each lifestyle change option, such as teleworking every day. ${\text{S}}{{\text{I}}_{i,k}}$ and ${\text{S}}{{\text{C}}_{i,k}}$ can be set as negative values to reflect rebound effects, the unintended increase in carbon footprints due to technology or behavioral changes. Substitution of consumption occurs when a consumption mode is shifted to a low-carbon mode, such as a shift from private cars to public transport, as calculated below:

$$\begin{align}{\text{HC}}_{i,j,k}^{{\text{sub}}} = \left\{ \begin{array}{ll} \displaystyle \sum\limits_{i^{\prime} \in {S_k}} {\left( {{\text{HC}}_{i^{\prime},j}^b \times {\text{S}}{{\text{C}}_{i^{\prime},k}} \times {R_{j,k}}} \right)} & \left( {i = {t_k}} \right) \\ \displaystyle 0 & \left( {i \ne {t_k}} \right). \end{array} \right .\end{align} \tag{ 11 }$$

The substituted amount of consumption ${\text{HC}}_{i,j,k}^{{\text{sub}}}$ by lifestyle change option k is the sum of the reduced consumption of items $i^{\prime}$ belonging to group ${S_k}$ that are the targets of the substitution to be reduced. The substituted amount is only added to the specific item ${t_k}$ targeted by the substitution to be increased. To illustrate, the maximum impacts from eating balanced food guide meals are modeled by first estimating the reduced amount of consumption of food items such as meat and cereals, and then adding the substituted amount of other food items such as vegetables and fruits. For lifestyle change options not involving substitutions, ${S_k}$ is set as empty and thus ${\text{HC}}_{i,j,k}^{{\text{sub}}}$ is zero.

Adoption rates ${R_{j,k}}$ indicate the level of uptake of lifestyle option k in city j, with values between 0 to 1 representing the extent of the lifestyle changes relative to the baseline consumption patterns. An ${R_{j,k}}$ value of 1 (100%) indicates the theoretical maximum uptake, or 'full adoption.' R_j,k values less than 1 indicate 'partial adoption'; e.g. teleworking only a few times per week. After estimating city-specific maximum impacts associated with the uptake of each lifestyle change option, city-specific partial adoption impacts can be estimated by establishing the city- and option-specific adoption rates.

With this methodology, the impacts of lifestyle change options can be estimated in multiple cities within a consistent framework. The lifestyle change options analyzed in the case study include the avoidance of some consumption, shifts between consumption items, and the adoption of improved products and services by households. These are limited to options that are currently available and can be adopted by households in urban context of the country. Given that the purpose of this study is to examine the potential of demand-side changes, the analyzed options include end-use technology options, such as electric vehicles and zero energy houses, but exclude supply-side technology options, such as renewable energy and efficiency improvements in the production processes. In the case study, lifestyle change options were identified through a review of the existing literature on the impact of consumer lifestyle changes and an examination of carbon footprint hotspots based on the estimated footprint data. In total, 65 options related to mobility, housing, food, goods, and leisure were identified to cover the most impactful options and major carbon footprint hotspots (table 1). (Data sources, assumptions, and the criteria for the identification of options are given in section 2 (tables S6–S9) in SI 1.) The city-specific full adoption carbon footprint reduction impacts of the identified lifestyle change options were quantified for the selected 52 cities. Embodied rebound effects (Sorrell 2012) (i.e. expected increases in carbon footprint through embedded emissions from the installment of equipment) were considered by partly increasing scenario GHG intensity ${\text{HI}}_{i,j,k}^s$ or consumption amounts ${\text{HC}}_{i,j,k}^s$ by setting negative values for reduction factors ${\text{S}}{{\text{I}}_{i,k}}$ or ${\text{S}}{{\text{C}}_{i,k}}$ in equations (7) and (8).

Due to the overlap of some options, the aggregated impact of adopting multiple lifestyle change options is not equal to the sum of the impacts of the individual options. Consequently, the city-specific aggregated impact of a combination of multiple lifestyle change options is quantified with the set of adoption rates for all options as follows:

$$\begin{align}{\text{IMP}}_j^s = {\text{ CF}}_j^s - {\text{CF}}_j^b\end{align} \tag{ 12 }$$

$$\begin{align}{\text{CF}}_j^s = { }\sum\limits_i \left( {{\text{HI}}_{i,j}^s \times {\text{HC}}_{i,j}^s} \right).\end{align} \tag{ 13 }$$

Here, the aggregated impact of multiple lifestyle change options ${\text{IMP}}_j^s$ in city j is given by comparing the baseline carbon footprints ${\text{CF}}_j^b$ in equation (5) and the scenario carbon footprints ${\text{CF}}_j^s$, which are determined by the scenario GHG intensities ${\text{HI}}_{i,j}^s{ }$ and consumption amounts ${\text{HC}}_{i,j}^s$ as follows:

$$\begin{align}{\text{HI}}_{i,j}^s = {\text{ HI}}_i^b \times \prod\limits_g \left( {1 - {\text{S}}{{\text{I}}_{i,g}}} \right)\end{align} \tag{ 14 }$$

$$\begin{align}{\text{HC}}_{i,j}^s = {\text{HC}}_{i,j}^b \times \prod\limits_g \left( {1 - {\text{S}}{{\text{C}}_{i,g}}} \right) + {\text{HC}}_{i,j}^s.\end{align} \tag{ 15 }$$

The scenario GHG intensities ${\text{HI}}_{i,j}^s{ }$ and consumption amounts ${\text{HC}}_{i,j}^s$ are given as the infinite product of the subtotal intensity reduction factors ${\text{S}}{{\text{I}}_{i,g}}$ or amount reduction factors ${\text{S}}{{\text{C}}_{i,g}}$ for option group g, adding the increased consumption of some items for substitution ${\text{HC}}_{i,j}^{{\text{sub}}}$. These infinite products consider the vertical overlap of options that are related to the same item of consumption but can be adopted together, such as the reduced impact of the shift from cars to public transport due to the shorter driving distance resulting from an earlier introduction of teleworking. The subtotal intensity reduction factors ${\text{S}}{{\text{I}}_{i,g}}$ and amount reduction factors ${\text{S}}{{\text{C}}_{i,g}}$ are given as follows:

$$\begin{align}{\text{S}}{{\text{I}}_{i,g}} = \sum\limits_{k \in {G_g}} ({\text{S}}{{\text{I}}_{i,k}} \times {R_{j,k}})\end{align} \tag{ 16 }$$

$$\begin{align}{\text{S}}{{\text{C}}_{i,g}} = \sum\limits_{k \in {G_g}} ({\text{S}}{{\text{C}}_{i,k}} \times {R_{j,k}})\end{align} \tag{ 17 }$$

$$\begin{align}{R_{j,g}} = \mathop \sum \limits_{k \in {G_g}} {R_{j,k}} \leqslant 1\end{align} \tag{ 18 }$$

Here, the reduction factors for option group g are calculated as the sum of the reduction factors (${\text{S}}{{\text{I}}_{i,k}}$ or ${\text{S}}{{\text{C}}_{i,k}}$) and adoption rates ${R_{j,k}}$ for the same group g under the conditions specified in (9) and (10). All lifestyle options are classified into option groups, designated g, in order to consider the horizontal overlap between options that cannot be adopted together, such as electric vehicles and hybrid vehicles. Options that do not have such horizontal overlaps form an individual group with only one member, and thus there is only one term in the summation. The condition specified in (18) refers to the maximum sum of the adoption rates ${R_{j,g}}$ in the same group g, not exceeding 100%. The increased consumption amount of an item i due to substitution ${\text{HC}}_{i,j}^{{\text{sub}}}$ is calculated as the sum of the increased consumption by substitution ${\text{HC}}_{i,j,k}^{{\text{sub}}}$ for lifestyle change option k (equation (11)) over all options. That is,

$$\begin{align}{\text{HC}}_{i,j}^{{\text{sub}}} = { }\sum\limits_k {\text{HC}}_{i,j,k}^{{\text{sub}}}.\end{align} \tag{ 19 }$$

With this methodology, the city-specific carbon footprint reduction pathways through lifestyle changes can be modeled and explored with different sets of adoption rates. In the case study, illustrative pathways for the selected 52 cities were developed assuming unified adoption rates across lifestyle change strategies for each city to determine the level of change required and the potential contribution of lifestyle changes to meet a given decarbonization target. The consumption-based per-capita target ${\text{C}}{{\text{F}}^{{\text{target}}}}$ is set as the upper 2030 target corresponding to the 1.5 °C target of the Paris Agreement adopted from an existing study (IGES et al 2019, Koide et al 2019). In addition, scenario analyses with lifestyle changes based on an efficiency strategy (i.e. the introduction of low-carbon technologies or the more efficient use of the same consumption mode) and a sufficiency strategy (i.e. a lessening of the physical or monetary amount of consumption, shifting modes of consumption, or behavioral changes that do not primarily rely on new technologies) were conducted for the 52 cities to examine the implications of different choices of lifestyle change options. While the case study analysis does not consider economy-wide rebound effects and employs a simplified assumption regarding adoption rates in the development of the illustrative scenarios, it provides useful insight into the potential of the methodology. (For a detailed discussion of the methodology and the limitations of the case study, please refer to sections 3 and 4 in SI 1.)

Figure 1 shows the estimated carbon footprints of household consumption in the targeted Japanese cities in the casestudy. The total carbon footprint of household consumption per capita was estimated to be 7310 ± 490 kgCO₂e/cap⋅yr (mean ± standard deviation) as of 2015. The estimated mean carbon footprints in the present study are comparable to those in a previous study of household carbon footprints in the country (7510 kgCO₂e in 2004) (Koide et al 2019); the slight differences can be explained by the different reference years and data sources for carbon intensity that were used in the studies. The footprints range from 8430 kgCO₂e (in 1. Mito city (the number here refers to the rank of the city relative to the baseline footprint as shown in figure 1)) to 5780 kgCO₂e (in 52. Naha city), a difference of 2650 kgCO₂e. Relatively large variations across cities are observed for mobility, housing, goods, and leisure (e.g. 1430 ± 290 kgCO₂e for mobility). Smaller variations are observed for food and services (e.g. 1300 ± 90 kgCO₂e in food). The range of the carbon footprints of Japanese cities estimated in the current study (5780–8430 kgCO₂e/cap⋅yr in 2015) differs from that in a prior study based on household survey data (3410–5000 kgCO₂e/cap⋅yr in 2011) (Jiang et al 2020) due to the proposed adjustment for the underreporting bias in the bottom-up survey data used in the current study.

Zoom In Zoom Out Reset image size

According to the result, on average, the carbon footprints need to be reduced at least by 56% to comply with the aspirational target of the Paris Agreement for limiting the global temperature increase to 1.5 °C. The required reduction level varies, ranging from a maximum of 62% (1. Mito city) to a minimum of 45% (52. Naha city). More detailed statistics for the estimated city-level footprints are available in table S11 (SI 1). The hotspots of the carbon footprints for the food, mobility, housing, and 'other' domains are shown in figure S1 for the country average and in figures S3 and S4 for selected cities.

Figure 2 shows the city-specific carbon footprint reduction impacts of the 65 lifestyle options in the case of full adoption (R_j,k = 1). As can be seen here, there is substantial variation among cities, particularly for car-related options in mobility, eco house and renewable energy options in housing, dietary and protein shifts in food, and leisure- and goods-related options. For example, electric vehicles charged by renewable energy have very heterogeneous footprint reduction impacts, ranging from −150 kg to −760 kgCO₂e/cap⋅yr (differing by a factor 5.0). Similarly, the impacts of rooftop solar panels with induction heating cookers vary from −1020 kg to −2160 kgCO₂e (factor 2.1). In general, the city-by-city estimated carbon footprint reduction impacts from major lifestyle change options in the present study are comparable to estimates reported in a previous country-level study focused on Japan (IGES et al 2019) (e.g. 140–620 kg vs. 520 kgCO₂e for electric vehicle, 1300–2100 kg vs. 960 kgCO₂e for renewable grid electricity and rooftop solar PV, and 220–420 kg vs. 340 kgCO₂e for vegetarian/vegan diet). The relatively small differences can be explained by the use of different data sources and the targeting of different areas.

Zoom In Zoom Out Reset image size

Such city-level variations indicate that the priority of the various lifestyle changes differs among cities. To illustrate, the ranges of impacts from mobility options such as ridesharing (−190 kg to −850 kgCO₂e) and teleworking (–160 kg to −440 kgCO₂e) overlap with the ranges of goods, leisure, housing, and food-related options, such as longer use of clothes (–120 kg to −280 kgCO₂e), community-based recreation (−120 kg to −440 kgCO₂e), wearing warm clothes (−70 kg to −280 kgCO₂e), and eating meat alternatives (–140 kg to −230 kgCO₂e/cap⋅yr). Notably, however, some of the lifestyle change options have consistently higher footprint reduction impacts than others, even considering between-city differences. For example, dietary and protein shifts (at least −280 kgCO₂e from a vegan diet, −140 kgCO₂e from meat alternatives) are far more impactful than efficient food sourcing and loss reduction (at most −10 kgCO₂e by local food, −50 kgCO₂e by seasonal food, −80 kgCO₂e/cap⋅yr by food loss reduction). Similarly, living in a zero-energy house has a four-digit impact (at least −1450 kgCO₂e), which is far larger than energy-saving behavior (at most, −280 kgCO₂e by wearing warm/cold clothes, −80 kgCO₂e from energy saving by nudging). Many of the more impactful options common among cities (e.g. electric vehicles, shift to public transport, renewable electricity, compact house, vegan and healthy diet) identified in the present study are compatible with the prominent options indicated in a 2020 review (Ivanova et al 2020). Estimated carbon footprint reduction impacts from lifestyle change options in the 52 cities, together with more complete statistics are provided in supplementary information 2 (hereinafter, SI 2) and table S13 (SI 1), respectively.

3.3.1. Illustrative Pathways to the Decarbonization Target

Illustrative carbon footprint reduction pathways to meet the consumption-based per-capita target for limiting the global temperature rise to 1.5 °C were developed for the 52 selected Japanese cities (figure 3). According to the results, ambitious urban lifestyle changes can shift carbon footprints to meet the decarbonization target, but the level of change required and the contribution of consumption domains significantly differ by city, with adoption rates for the lifestyle changes needed to meet the target ranging from 62% to 87%. These percentages are close to those found in a previous study at the country level, which concluded that a 65%–75% adoption rate for the various options would be necessary to meet the 2030 target (IGES et al 2019). The required coverage of demand-side options is also comparable to a previous study, which presented scenarios for a 90% reduction of carbon footprints in London (Hersey et al 2009).

Zoom In Zoom Out Reset image size

The carbon footprint reduction pathways can be understood using two metrics: reduction of the consumption amount, and reduction in the carbon intensity. Figure 4 illustrates how urban consumers' carbon footprint hotspots are addressed through lifestyle changes in some selected cities (cities with the largest and smallest baseline footprints in each consumption domain). In the illustrative pathways, reductions in both consumption amounts (e.g. energy use, mobility distance, food consumption, and goods and service expenditure) and their intensities are reduced through a variety of urban lifestyle change options. Importantly, the level of reduction in both amount and intensity varies between cities, due to their different adoption of lifestyle change options to address carbon footprint hotspots.

Zoom In Zoom Out Reset image size

The footprint reductions in housing range from −2.2 tCO₂e (17. Fukui city) to −1.0 tCO₂e/cap⋅yr (52. Naha city). In the illustrative pathways, electrification and the shift to renewable electricity and energy demand reduction including through zero-energy and compact houses with energy saving behavior are assumed. Similarly, the impacts in the mobility footprints range from −2.0 tCO₂e (1. Mito city) to −0.70 tCO₂e/cap⋅yr (50. Osaka city). Regarding mobility, distance reduction through teleworking, more compact cities, and micro leisure, a shift to public transport and bicycles, and sharing and the electrification of cars are considered.

In food (cooking at home and ready-made meals), the reduced footprints range from −0.50 tCO₂e (19. Kitakyushu city) to −0.24 tCO₂e/cap⋅yr (52. Naha city). Here, a shift in the amount of consumption to beans and vegetables, a reduction in the GHG intensity of food items through changes in the composition of food and field-cultured, locally-produced vegetables are included. In the category of 'others' (a combination of leisure, goods, and services), the reduced footprints range from −1.2 tCO₂e (9. Kawasaki city) to −0.44 tCO₂e/cap⋅yr (41. Aomori city). In the pathways, reduction in the consumption of clothes, hobby goods, and consumables through longer use and material savings is considered. These results show that even with the simplified assumption of unified adoption rates per city, the shape of future lifestyles and hotspots significantly varies between cities. Table S12, figures S2, S5 and S6 (SI 1) provide additional details.

3.3.2. Scenario Analysis of Efficiency and Sufficiency Lifestyle Changes

Figure 5 gives the results of the scenario analyses using different combinations of efficiency and sufficiency lifestyle changes in the 52 cities. Importantly, it was found that maximum adoption of either efficiency or sufficiency lifestyle changes alone will not meet the 1.5 °C target. A 100% uptake of all efficiency options (e.g. life cycle carbon minus houses, electric vehicles charged with renewable energy, ridesharing, local and seasonal food with food loss reduction) reduces the total household carbon footprint by −39% to −50%, resulting in per-capita footprints of 3.5 t to 4.2 tCO₂e, all of which exceed the target. Similarly, the uptake of all sufficiency options (e.g. telework, micro tourism, a modal shift to bus, train or bicycle, energy saving behavior, vegan diets and healthy lifestyles, longer use of durables, and community-based leisure) reduces footprints by −42% to −47%, resulting in per-capita footprints of 3.4 t to 4.4 tCO₂e, which, again, are all above the target.

Zoom In Zoom Out Reset image size

The scenario analysis indicated that an ambitious uptake of both efficiency and sufficiency lifestyle changes will be needed to decarbonize urban consumption. For example, if a 100% adoption of efficiency lifestyle changes is assumed, including an additional 25%–75% of the sufficiency options will be required in order to reduce the footprints to the targeted level. Similarly, if a 100% uptake of the sufficiency options is assumed, 25%–75% of the efficiency options would be required. More detail is available in figure S7 (SI 1).