Development patterns, material metabolism, and greenhouse gas emissions of high-speed railway in China

High-speed railway (HSR) is a key city-to-city transport infrastructure providing efficient mobility. China has built the largest HSR system, which enhances commuting and economic development but entails substantial material consumption and greenhouse gas (GHG) emissions. Traditional environmental analyses of rail have focused on operational emissions, highlighting rail’s transport energy advantages, yet there are tradeoffs between pre-use (embodied) and in-use emissions: higher HSR speeds may require more energy-intensive construction materials. China plans an “eight verticals and eight horizontals” HSR network by 2035 amid rapid urbanization and growing intercity demand, raising sustainability concerns from material requirements and embodied emissions. Existing studies of HSR material metabolism are limited—often line-specific or focused only on stocks—leaving national-scale stock-flow characterization and carbon replacement value (CRV) largely unexplored. Leveraging emerging spatial data and bottom-up modeling, the study asks: (1) What are the development patterns and material metabolism characteristics in China’s HSR? (2) What are the associated GHG emissions and environmental impacts of HSR development?

Prior research has emphasized operational energy and emissions of rail and the transport benefits of rail’s energy efficiency, but the interplay between embodied and operational emissions with increasing speed is less quantified. Material metabolism approaches have been applied to buildings and materials (e.g., steel, cement), with few applications to HSR, typically at single-line scales or limited to stock estimates. National-scale, spatiotemporally explicit assessments of HSR stocks and flows are scarce, and CRV for HSR infrastructure has been under-discussed. Big data and high-resolution mapping have advanced built environment stock analysis but have not been fully utilized for HSR. Comparative assessments across passenger transport modes (HSR, road, aviation) for stocks, embodied emissions, and operations are also limited in scope.

The study compiles a national HSR database (2008–2020) covering 152 lines and 1,064 stations, with attributes including length, bridge/tunnel/track/culvert lengths, speed, geolocation, opening dates, and station volumes/classes. Data sources include national statistics, official company and government reports, on-site surveys and interviews, and web-crawled geospatial data (primarily Baidu Map) rectified for accuracy. Material composition indicators (MCIs) were compiled from bills of quantities for >50 HSR projects, aggregating >400 material use processes into seven end-use components (bridge, subgrade, tunnel, track, station, culvert, and EPCS systems) and 19 material types, later grouped into seven bulk material categories (cement, fly ash, gravel, sand, steel, water, others). Road and aviation MCIs were similarly compiled. Bottom-up stock estimation integrates physical sizes of components with MCIs: MS_{m,i} = Σ(PS_i × MCI_{m,i}). Material flows (inflows for new construction and outflows/wastes) are derived via mass balance between consecutive years, adding maintenance flows (non-structural within 2008–2035 given 100-year design life for major structures) and construction-related losses (temporary facilities, formwork). Spatial analysis maps line polylines and station polygons with stock attributes to compute provincial cumulative stocks and stock densities (per area, per capita) and Gini coefficients (2010, 2020, 2035). Environmental impacts include CRV calculated from cradle-to-gate emission factors (Chinese Life Cycle Database and literature) via CRV_{m,i} = EF_m × MS_{m,i}, and operational emissions estimated by OE_t = Σ(OEF_t × PT_t) using passenger turnover and mode-specific operational emission factors (HSR electricity, road fuels, aviation fuel). Uncertainty analysis applies Monte Carlo simulation (10,000 runs, 95% confidence intervals) on key parameters (physical sizes, MCIs, emission factors, passenger turnover) to quantify ranges for stocks and CRV; sensitivity analyses identify influential years, materials, and components. Software: ArcGIS 10.2 and eBalance 4.7.

- Network growth: HSR expanded from 1,039 km and 24 stations (2008) to 38,914 km and 1,064 stations across >150 lines (2020). Peak growth occurred 2010–2015 with 39% annual length increase, then 11% annually 2016–2020. China’s HSR length accounts for ~60% of global total. - Material stocks: Cumulative HSR stocks rose from 0.13 Gt (2008) to 3.65 Gt in 2020 (mean; 3.60–3.71 Gt) and are projected to reach 4.9 Gt by 2035. Annual net additions: 71 Mt (2008), ~342 Mt/yr (2011–2015), ~277 Mt/yr (2016–2020). - Composition and components (2020): Nonmetallic minerals dominate in-use stocks (88%): gravel 1,756 Mt (48%), sand 891 Mt (24%), cement 560 Mt (15%). By end-use, bridges 37.2% (1.36 Gt) and subgrade 34.4% (1.26 Gt) lead. Bridge stocks are mainly box girders (39%); subgrade stocks mainly slope protection (50%) and railway subgrade (33%). Bridge-tunnel ratio increased, raising tunnel shares (tunnel stocks up 11× since 2008). - Speed and technology effects: Average material stock density fell from 110 t/m (2008) to 92 t/m (2020) while average speed rose from 240 to 276 km/h. Higher-speed design increases specific component densities (e.g., steel rail density 0.24 t/m at 200 km/h to 0.28 t/m at 350 km/h; subgrade 36.3→37.4 t/m; tunnel 73.4→77.4 t/m). Technology advances offset demands: adoption of 40 m box girders can save 20–39% materials vs 24–32 m spans and could avoid 292 Mt if scaled nationally. Ballastless track reduces track base material by ~26% (3.5 vs 4.7 t/m). Despite material efficiency, faster HSR lines emit 10–50% more operational carbon than 200–250 km/h lines. - Spatial disparities: Stocks shifted toward balance across regions. Western regions’ cumulative stocks grew from 39 Mt (7%) in 2010 to 840 Mt (22.5%) in 2020 and projected 1,283 Mt (26.4%) by 2035. Gini of provincial cumulative stocks declines from 0.56 (2010) to 0.29 (2035); per area and per capita Gini decline from 0.68/0.62 to 0.39/0.33. However, pronounced Hu-Line disparity persists (2020: 9.7% NW vs 90.3% SE). Per area extremes (2020): Shanghai 2,082 t/m² vs Qinghai 35 t/m² and Xinjiang 47 t/m². Per capita extremes: Hainan 6,453 t/cap; Gansu 5,844 t/cap; Ningxia 409 t/cap; Shanghai 531 t/cap. - Material flows (to 2020): Production for HSR construction includes gravel 1,744 Mt, sand 896 Mt, cement 561 Mt, steel 133 Mt, water 206 Mt, fly ash 92 Mt. Concrete engineering demands 1,413 Mt of key inputs; only 54 Mt steel is used for track. Waste generated is modest (26.3 Mt by 2020), mainly from temporary facilities; 0.68 Mt steel from formwork recycled, most concrete waste landfilled/backfilled. - Emissions: HSR CRV reaches ~1,008 Mt by 2020 (uncertainty 942–1,031 Mt); energy-intensive materials dominate CRV (cement 49%, steel 47%). HSR operational emissions are ~31 Mt/yr (2020). Across intercity passenger transport (2020): total operational emissions 143 Mt (aviation 71 Mt, HSR 31 Mt, road 40 Mt). Stocks per passenger-km (2020): road 175 kg/pkm, HSR 8 kg/pkm, aviation 0.8 kg/pkm. HSR CRV per pkm declined from 19 kg/pkm (2008) to 2 kg/pkm (2020). Operational emissions per pkm (2020): aviation 112 g/pkm, road 87 g/pkm, HSR 65 g/pkm. - Resilience effects: COVID-19 reduced aviation and road passenger turnover roughly twofold; HSR turnover decreased marginally, increasing per passenger operational intensities more for road/aviation.

The study demonstrates that China’s rapid HSR expansion has created large in-use material stocks dominated by aggregates and cement, yet technological advances have reduced material intensity over time even as average operating speeds increased. Spatial inequalities in HSR infrastructure are narrowing, particularly with expansion into western provinces, although significant disparities remain along the Hu-Line and in per-area accessibility. Material flow analysis highlights heavy reliance on sand and gravel, pointing to potential ecological pressures and the need for aggregate-saving designs and substitutes. From a climate perspective, HSR exhibits lower operational emissions per passenger-km than road and aviation and comparatively low construction material intensity per service delivered, supporting its role in green transport transitions. However, embodied emissions remain substantial due to cement and steel, emphasizing the importance of deploying low-carbon production technologies (e.g., hydrogen-based steel, CCS for cement) and material efficiency measures (long-span girders, ballastless tracks). Comparative analysis across modes underscores broader decarbonization levers: cleaner power systems for HSR, electrification for road transport, and power-to-liquid fuels for aviation. Enhancing HSR ridership via pricing and logistical strategies can further improve system-wide efficiency and reduce per passenger emissions.

This work provides a comprehensive, bottom-up, spatiotemporal assessment of China’s HSR development, quantifying stocks, flows, and GHG emissions from 2008 to 2035. HSR stocks increased to 3.65 Gt by 2020 and are projected to reach 4.9 Gt by 2035, with nonmetallic minerals dominating. Technological innovations have reduced material intensity and can deliver significant resource savings if widely adopted. Spatial analyses reveal narrowing provincial disparities but persistent accessibility gaps. The HSR CRV is approximately 1.01 Gt by 2020, and operational emissions are 31 Mt/y; per passenger-km, HSR outperforms road and aviation on operational emissions and material intensity. The findings inform green transition strategies: accelerate low-carbon steel and cement, expand material efficiency (long-span girders, ballastless track), develop aggregate substitutes, and decarbonize electricity. Future research should extend to broader (continental/global) scales, refine line- and component-level MCIs, improve spatial resolution of stock mapping, and integrate life-cycle operational dynamics under evolving energy systems and ridership patterns.

Results rely on compiled physical sizes, MCIs, and GHG emission factors that vary spatially and temporally due to construction conditions and technology advances. Maintenance was assumed not to trigger large structural replacements within 2008–2035. Uncertainty was quantified via Monte Carlo simulation (10,000 runs, 95% CI), yielding 2020 HSR stocks of 3,601–3,713 Mt and CRV of 942–1,031 Mt. Sensitivity analyses indicate stocks are most influenced by specific years (2014, 2018, 2015) and components (subgrade, track, bridge); CRV uncertainty is dominated by cement and steel emission factors. The study could be expanded to larger geographic scopes, with more detailed, line-specific MCIs and higher spatial resolution of components. Operational emission factors and passenger turnover trends also contribute uncertainty, especially under disruptions like COVID-19.