Mitigation synergy and policy implications in urban transport sector: a case study of Xiamen, China

The LEAP model is applied extensively for setting energy-related carbon emission scenarios (Cai et al 2013, Wu et al 2021). It is basic framework is shown in figure S2. Recently, the LEAP model has been extended to energy-related AP emission projections (He et al 2010, Jiang et al 2021). However, there are still incomplete AP emissions in urban transport sector because the emission factors of fuel by different emission standards are not embodied. Hence, we establish a customized LEAP model for GHGs and nine APs in Xiamen's transport sector by adopting local AP emission factors and referring to the AP emission inventory for Chinese cities (MEPC 2017). Our study period established in the LEAP model from 2013 to 2060. The emissions from 2013 to 2020 are considered historical, with 2020 serving as the baseline year for future predictions.

As seen in table 1, we first unify and establish sector classification, including four sector categories (i.e. road, shipping, aviation and rail) and sixteen sub-sectors, to make GHG and AP emissions comparable. The detailed classification in the LEAP model is shown in figure S3. The estimates of GHG emissions from urban transport sector include three types (i.e. CO₂, CH₄ and N₂O), and the final results are presented by GHGs due to the small proportion of CH₄ and N₂O. Nine APs (i.e. SO₂, NO_X , CO, PM₁₀, PM_2.5, BC, OC, VOCs and NH₃) are individually analyzed. For GHGs, an energy consumption-based method based on the IPCC guidelines is used to estimate historical GHG emissions, as shown in equation (1)

$$\begin{equation}C{E_{i,j,k}} = \sum\nolimits_{i,j,k} {\left( {{A_{i,j,k}} \times E{F_{i,j,k}} \times GW{P_i}} \right)} \end{equation} \tag{ 1 }$$

where ${\text{C}}{{\text{E}}_{i,j,k}}$ denotes the emissions of a specific GHG type i on energy j (i.e. gasoline, diesel, natural gas, fuel oil, jet kerosene, electricity, hydrogen and sustainable aviation fuel) consumed in sector/sub-sector k. ${{\text{A}}_{i,j,k}}$ denotes the corresponding energy consumption based on a bottom-up survey provided by local government departments. ${\text{E}}{{\text{F}}_{i,j,k}}$ denotes the associated factors obtained from the IPCC guidelines (IPCC 2006). ${\text{GW}}{{\text{P}}_i}$ denotes the global warming potential of GHG type i referring to the IPCC fifth assessment report (IPCC 2014), that is, 1, 28 and 265 for CO₂, CH₄ and N₂O, respectively.

Except for road sector, APs include eight types (i.e. SO₂, NO_X , CO, PM₁₀, PM_2.5, BC, OC and VOCs). We apply a dynamic technology-based approach to the multi-year AP emission inventory following previous studies (Streets et al 2006, Bian et al 2019), as shown in equation (2). Specifically, the emission factor of SO₂ is estimated by equation (3), and S represents the sulfur content in fuel

$$\begin{equation}A{E_{m,n,k}} = \sum\nolimits_{m,n,k} {[{A_{m,n,k}} \times E{F_{m,n,k}} \times \left( {1 - {{{\eta }}_{m,n,k}}} \right)]} \end{equation} \tag{ 2 }$$

$$\begin{equation}{\text{E}}{{\text{F}}_{{\text{S}}{{\text{O}}_2}}} = 2 \times S\end{equation} \tag{ 3 }$$

where $A{E_{m,n,k}}$ and ${A_{m,n,k}}$ denote the emissions and activity level n (i.e. energy consumption and flight movements) of a specific AP type m in emission sector/sub-sector k. Activity-level data are from government departments and the yearbook of Xiamen's special economic zone (XMBS 2014–2014-2021). $\,{{\text{EF}}_{i,j,k}}$ denotes emission factors based on the guidelines for the development of an AP emission inventory for Chinese cities (MEPC 2017). ${\eta _{i,j,k}}$ denotes the removal efficiency due to technological penetration and emission controls.

Additionally, road sector emissions (RE) are calculated by vehicle populations (P), annual average vehicle kilometers traveled (VKT) and vehicle emission factors (EF) using equation (4). Here, m represents a specific AP in nine types; p and q represent the vehicle type and emission standard, respectively. Data are from the yearbook of Xiamen's special economic zone and field investigation (XMBS 2014–2014-2021)

$$\begin{equation}R{E_{m,p,q}} = \sum\nolimits_{m,p,q} {({P_{m,p,q}} \times VK{T_{m,p,q}} \times E{F_{m,p,q}})}.\end{equation} \tag{ 4 }$$

To quantify the effect and degree of mitigation synergy for GHG and AP emissions under different clean air measures, the coordinated control system and pollutant reduction cross elasticity (Els) are employed in Xiamen's transport sector during 2013–2020. These are two methods ordinarily used in previous studies (Alimujiang et al 2020). Clean air measures during 2013–2020 are summarized in table 2.

A coordinated control system is a two-dimensional coordinate system that intuitively manifests the co-benefits of GHGs and APs. We use the abscissa and ordinate to represent cumulative emission reductions of GHGs and APs, respectively, during this study period. Under different clean air measures, a specific AP and GHG abatement can fall on a point in the coordinate system (figure 1(a)). For a given constant level of AP emissions reductions, a smaller angle α indicates a better effect on GHG mitigations.

Zoom In Zoom Out Reset image size

Furthermore, we utilize Els to evaluate which matters between GHGs and APs have a larger abatement degree under co-mitigation measures based on the range of Els. The Els estimates are obtained by equation (5). Additionally, figure 1(b) displays the positive/negative synergy and co-reduction degree

$$\begin{equation}{\text{El}}{{\text{s}}_{{\text{GHGs}}/{\text{AP}}}} = \frac{{\Delta {\varphi _{{\text{GHGs}}}}/{\varphi _{{\text{GHGs}}}}}}{{\Delta {\varphi _{{\text{AP}}}}/{\varphi _{{\text{AP}}}}}}.\end{equation} \tag{ 5 }$$

Here, ${\text{El}}{{\text{s}}_{{\text{GHGs}}/{\text{AP}}}}$ is the cross-elasticity of a specific AP with GHGs, and $\Delta {\varphi _{{\text{GHGs}}}}/{\varphi _{{\text{GHGs}}}}$ and $\Delta {\varphi _{{\text{AP}}}}/{\varphi _{{\text{AP}}}}$ represent the GHG and AP change rates, respectively.

Based on existing emission policies and regulations released by Chinese and local governments, the emission scenarios, which are associated with control measures executed during 2013–2060, are set in the aforementioned LEAP model, assessing mitigation synergy effects and potential. There are seven emission scenarios, namely, the business as usual (BAU), eliminate and renew transportation (ERT), guide green travel (GGT), upgrade oil quality (UOQ), improve vehicular emission standards (IVE), adjust energy structure (AES) and optimize transport structure (OTS) scenarios (all abbreviations are listed in table S1). The scenario descriptions are shown in table 2, and the main references are summarized in table S2. Considering the timeline for scenario implementation (table 2), it can be observed that all scenarios cover the 2013–2020 period, and four scenarios (i.e. BAU, GGT, AES, and OTS) are considered from 2021 to 2060. The key parameters set in the LEAP model are shown in table S3.

GHG and AP emissions of Xiamen's transport sector from 2013 to 2020 are estimated. Figure 2 shown both GHG and AP emissions have similarities and differences. For GHGs, emissions increased steadily from 5.4 Mt in 2013 to 8.6 Mt in 2019, yet GHG emissions dropped to 8.2 Mt in 2020 due to the influence of the COVID-19 pandemic, especially GHG emissions from international flights in 2020, which decreased by approximately 15% compared to 2019. Additionally, emissions from taxis and buses have also declined in recent years and varied considerably in 2020, which benefits from the vehicular fuel transition from oil to electricity and natural gas. Light-duty vehicles are the dominant sector driving the emission variations, and their contribution accounted for 31% in 2013 and increased to 46% in 2020.

Zoom In Zoom Out Reset image size

Similar to GHGs, emissions from light-duty vehicles are also important for CO, VOCs and NH₃, especially for NH₃. NH₃ emissions originate from road sector and mostly come from light-duty vehicles, with contributions increasing from 64% in 2013 to 84% in 2020. The agricultural sector is the largest contributor to NH₃ emissions (Huang et al 2012), whereas its emissions are concentrated in rural areas, and vehicular NH₃ emissions elevate with urbanization and increasing transport populations. With a high urbanization of 90% in 2021 (XMBS 2022), Xiamen's NH₃ emissions from the transport sector were the third-largest source (figure S4). Thus, it is imperative to continually pay attention to vehicular NH₃ emissions in Xiamen. Regarding CO and VOCs, these emissions increased from 2013 to 2015 and declined after 2015. For VOCs, VOC emissions increased from 10.0 kt in 2013 to 11.5 kt in 2015 and decreased to 9.7 kt in 2020. Motorcycles and light-duty vehicles are two major contributors to VOC emissions in the transport sector, with emissions decreasing after 2015 and 2017, respectively, as a result of the ban on motorcycles and the strengthening of vehicular emission standards.

Due to the lack of strict regulations on vehicular NH₃ abatement, its emissions climbed during 2013–2020. SO₂ emissions also increased from 1.3 kt in 2013 to 2.4 kt in 2020. Unlike NH₃, the reasons for the rising SO₂ emissions are attributed to shipping sector; river boats are the dominant SO₂ emitter, with a 161% increment. In addition to SO₂, shipping sector emissions are prominent and growing, with annual average contributions of NO_X , PM₁₀, PM_2.5, BC and OC of 28%, 45%, 46%, 48% and 48%, respectively, during the 2013–2020 period. Notably, although shipping sector GHG emissions are minor, their contributions grew from 9% in 2013 to 14% in 2020. Therefore, Xiamen should focus efforts on shipping sector emissions to yield co-benefits for reducing GHG and AP emissions, especially for SO₂, NO_x , PM₁₀, PM_2.5, BC and OC.

NO_X emissions in the transport sector increased from 35 kt in 2013 to 46 kt in 2018 and decreased to 44 kt in 2020. Heavy-duty trucks are the largest contributor to NO_X emissions, with an average contribution of 35%, followed by river boats (23%). NO_X emissions from buses have evidently dropped since 2017 at a 71% rate. The characteristics of PM₁₀, PM_2.5, BC and OC emissions are in accordance. For PM_2.5, emissions grew from 1.5 kt in 2013 to 2.1 kt in 2018 before a plateau. PM_2.5 emissions from heavy-duty trucks are also important, with an average contribution of 23%, second to river boats (39%). Similar to NO_X , PM_2.5 emissions from buses also decreased after 2017.

To evaluate clean air actions to co-reduce GHG emissions during 2013–2020, we estimate reductions in nine APs and GHG emissions from the road, shipping, aviation and rail sectors by comparing the cumulative difference between six scenarios (i.e. ERT, GGT, UOQ, IVE, AES and OTS) and BAU, as shown in figure 3(a). According to estimations, the implementation of clean air measures for Xiamen's transport sector were able to avoid emissions of GHGs, SO₂, NO_X, CO, PM₁₀, PM_2.5, BC, OC, VOCs and NH₃ with 5.3 Mt, 1.5 kt, 32.5 kt, 45.4 kt, 2.2 kt, 2.1 kt, 1.1 kt, 0.4 kt, 7.3 kt and 0.7 kt, respectively. Overall, road sector abatement is the most prominent, especially for CO, NO_X , VOCs and GHGs. In this sector, GGT and AES are major measures for mitigation synergy between GHGs and APs, and UOQ only helps reduce APs, especially NO_X , PM₁₀, PM_2.5 and BC. Following shipping sector emissions, OTS is the principal scenario for GHG and AP mitigation. The reduction in emissions from the aviation and rail sectors is small, and optimizing the transport structure, such as developing an air-freight service hub and combined rail-water transport, is the sole measure helping mitigation in the aviation and rail sectors.

Zoom In Zoom Out Reset image size

The degree to which each scenario contributes to mitigation synergy between GHGs and APs for the whole transport sector is presented in figure 3(b). No clean air measures in Xiamen's transport sector during the 2013–2020 period increased GHG emissions; instead, they mainly cut emissions. Two scenarios (namely UOQ and IVE) reduced AP emissions but did not or scarcely contribute to GHG abatement, where other four scenarios (i.e. OTS, ERT, GGT and AES) had co-effects on GHG and AP abatement to some extent. For VOCs, an emission reduction of approximately 1 Mt GHGs was induced by cutting VOC emissions by 5.5 kt, 1.4 kt, 1.2 kt and 1.0 kt under ERT, AES, OTS and GGT scenarios, respectively. Thus, GGT and OTS have the largest co-effects between VOC and GHG abatement. Note that emission reductions may not be linear between GHGs and APs under measures; here, only the degree of synergy under different scenarios is compared. Similarly, GGT has massive mitigation and GHG co-benefits for SO₂, NO_X and CO emissions; under the scenario, reducing 0.2 kt SO₂, 0.2 kt NO_X and 1.6 kt CO emissions co-drives 3 Mt GHG mitigation. Additionally, although there is only small mitigation of APs, significant mitigation is observed under some scenarios. For example, GGT has a small effect on cutting PM₁₀, PM_2.5, BC and OC emissions but is significant for GHG emission reductions. Thus, we consider them to have weak co-effects. Based on these results, AES has good mitigation synergy between AP (i.e. PM₁₀, PM_2.5, BC, OC and NH₃) and GHG emissions, co-reducing PM₁₀, PM_2.5, BC, OC and GHG emissions by 0.5 kt, 0.5 kt, 0.2 t, 0.2 t and 1.8 Mt, respectively. The GHG degrees of synergy from different clean air measures for nine APs are summarized in table 3.

Among four significant scenarios, i.e. OTS, ERT, GGT and AES, we use Els ranges to evaluate which one exhibited a higher degree of reduction between APs and GHGs under different scenarios during 2013–2020 (table 3). ERT has higher mitigation effects for APs than for GHGs (1 > Els > 0), OTS also has a greater degree of reduction, mostly for APs, except for CO and VOCs. These indicates two measures has more significant effects on most AP mitigation. Additionally, GGT and AES are more effective at reducing GHG emissions than most APs (Els > 1), highlighting the crucial role of clean air measures (particularly in GGT and AES) in driving GHG mitigation in the future, which is a key pathway for achieving future carbon emission reductions.

Future emissions of GHGs and APs for Xiamen's transport sector by scenarios during 2020–2060 are shown in figure 4. Detailed scenario descriptions are summarized in table 2. Under BAU scenario, if few intervening measures are adopted during 2021–2060, GHG emissions will increase by 55% from 2021 to 2038, and rise slightly during 2039–2060. SO₂, CO, VOC and NH₃ emissions will increase before decreasing with peaks in 2036, 2029, 2033 and 2037, respectively. Unlike their emissions increasing steadily until the peak year, NO_x emissions will have a momentary drop before climbing to a peak in 2037 and then falling. PM₁₀, PM_2.5, BC and OC emissions will also have momentary drops before 2025, after which they are expected to increase until 2060.

Zoom In Zoom Out Reset image size

Regarding OTS scenario, it seems clear that GHG, SO₂, NO_x , CO, VOC and NH₃ emissions will be lower than those under BAU scenario. GHG emissions will increase by 39% from 2020 to 2038 and then decline by 5% until 2060. CO, VOCs and NH₃ emission trends will be similar to those of the corresponding emissions under BAU scenario, with emission peaks in 2025, 2032 and 2036, respectively, earlier than those under BAU scenario. The SO₂ and NO_x emission trends will also be similar to those under BAU scenario before 2050, but edge up after 2050, mainly resulting from the slowing declines in the population of heavy-duty trucks due to 'road to water and rail'. These will be also reflected in the slower declines of CO, VOCs and NH₃ emissions. For PM₁₀, PM_2.5, BC and OC, their emission trends will be slightly lower than those under BAU but surpass those since 2051, mostly because shipping sector emissions will increase more than road sector emissions will decrease due to 'road to water'. Compared to emission pathways under BAU scenario, emissions under GGT scenario will decrease to some extent for GHGs, SO₂, CO, VOCs and NH₃. However, for NO_x , PM₁₀, PM_2.5, BC and OC, the emissions under GGT will mostly coincide with those under BAU scenario. This is mainly because the population of light-duty vehicles will be slightly affected by GGT scenario, and emission factors of NO_x , PM₁₀, PM_2.5, BC and OC will undergo significant declines over time under BAU scenario. Thus, mitigation effects under GGT scenario are expected to be minimal. Note that both GHG and AP emissions under AES scenario will have obvious mitigation. Similar to the emission pathways under OTS and GGT scenarios, GHG emission pathways under AES scenario will decrease by 27% after increasing by 79% with a peak in 2033. Additionally, SO₂ and NH₃ emissions will have similar trends, with emission peaks in 2030 and 2035, respectively. NO_x , CO, PM₁₀, PM_2.5, BC and OC present overall substantial declines.

Based on emission pathways under different scenarios, GHG and AP abatement will accumulate under different scenarios compared to BAU scenario during 2020–2060 (figure 5). Overall, there will be significant potential for emission reductions, particularly in the road and shipping sectors.

Zoom In Zoom Out Reset image size

Under OTS scenario, road sector is expected to experience a significant reduction in cumulative emissions, especially for GHG, SO₂, NO_x , CO, VOCs and NH₃, mainly due to 'road to water and rail'. Their cumulative emissions are projected to be 58 Mt, 10 kt, 334 kt, 167 kt, 16 kt and 5 kt, respectively. In contrast, cumulative emissions from the shipping and rail sectors will slightly increase. These results indicate that 'road to water and rail' will have good synergistic effects on GHG and SO₂, NO_x , CO, VOC and NH₃ mitigation in the transport sector. Note that the cumulative emission increments from the shipping and rail sectors will counteract the cumulative emission reductions from road sector for PM₁₀, PM_2.5, BC and OC, which is also reflected in figure 4, which shows that their emission pathways under OTS scenario will approach those under BAU scenario.

Compared to OTS scenario, cumulative emission mitigation under the AES scenario will come from multiple sectors. For example, the cumulative GHG emission mitigation will be simultaneously from the road, shipping, aviation and rail sectors, whose cumulative emission reductions will amount to 107 Mt, 31 Mt, 32 Mt and 4 Mt, respectively. Similar to OTS scenario, in addition to GHGs, road sector under the AES scenario will have a significant cumulative mitigation for SO₂, NO_x , CO, VOCs and NH₃, amounting to 7 kt, 80 kt, 222 kt, 45 kt and 13 kt, respectively, which will be attributed to road mitigation measures, such as strengthening transport electrification, improving the energy efficiency standards for fuel vehicles and popularizing clean energy and new energy vehicles. At the same time, for SO₂, NO_x , CO, PM₁₀, PM_2.5, BC, OC and VOCs, the shipping sector will be a large contributor, with cumulative emission reductions of 39 kt, 460 kt, 230 kt, 37 kt, 35 kt, 20 kt and 6 kt, 60 kt, respectively, which will be driven by promoting clean vessels and improving the energy efficiency standards for fuel vessels. Due to the promotion of sustainable aviation fuel, the cumulative emission reductions from the aviation sector will affect GHGs, NO_x , CO and VOCs, with decreases in NO_x , CO and VOCs will be 43 kt, 240 kt and 7 kt, respectively.

For GGT scenario, cumulative emission mitigation will come only from road sector, and GHG, SO₂, CO, VOCs and NH₃ will be significant, with reductions of 118 Mt, 7 kt, 91 kt, 233 kt, 54 kt and 17 kt, respectively, which will also be evident in their emission pathways under GGT scenario, which differ widely from those under BAU scenario (figure 4).