Carbon flow model tracks industrial park emissions with 94% accuracy

Industrial parks sit at an awkward middle scale in the fight against climate change. They are too large and too internally complex to be treated like a single building, yet too small and too behaviorally intricate to be captured by the citywide carbon inventories that most governments rely on. A new study argues that this mismatch has left planners without a reliable way to measure, attribute, and ultimately reduce the emissions generated inside park boundaries, and it proposes a detailed alternative built on carbon flow modeling and system dynamics.

The research, published in the journal Frontiers of Urban and Rural Planning, was conducted by Juan Sun and Tinghai Wu, who set out to construct a carbon flow model of industrial parks that clarifies emissions pathways, identifies the entities responsible for them, and supports planning-oriented carbon reduction strategies. Their central claim is that existing accounting methods designed for macro city scales or micro building scales fail to accommodate the complex systemic characteristics of parks, which combine residential, industrial, transport, energy, and ecological functions within a single managed territory.

To address this gap, the authors drew on carbon flow theory and system dynamics to build an integrated model comprising five major subsystems: residential, production, transportation, renewable energy production, and carbon sink. Within this framework, they distinguish between two categories of influencing factors. Fundamental factors are relatively fixed elements such as planned residential population, climate zoning, and regional electricity carbon emissions factors, which can be estimated directly from planning indicators and location. Dynamic factors, by contrast, reflect the operational status of the park, including building vacancy rates, the adoption rate of low-carbon technologies, and the intensity of daily activities. It is the interplay of these two factor types, the study argues, that determines how closely a calculated emissions figure matches reality.

The methodological core of the work is a translation exercise. The authors note that mainstream accounting approaches split into two camps: production-end inventories derived from IPCC guidelines, which categorize emissions by generation mechanism into energy, industrial processes, land use, and waste; and consumption-side models such as the Greenhouse Gas Protocol for Cities, which trace emissions to individual entities across product life cycles. Each has limitations when applied to a park. Production-side methods struggle to assign responsibility across diverse urban activities, while consumption-side methods emphasize single entities and lack comprehensive accounting for multiple systems within a specific spatial context. The proposed framework therefore integrates both perspectives, tracing emissions sources at the production end while clarifying the emissions impacts of various system activities at the consumption end.

From this translation, the study delineates twelve distinct carbon emissions responsible entities. The building sector, for example, is subdivided into energy-related emissions from public buildings, energy-related emissions from residential buildings, and emissions reductions from building energy efficiency. The transportation sector is categorized by travel subjects and purposes into commuting emissions from residents, non-commuting emissions from residents, and freight transport emissions. The industry sector is segmented by emissions stage into energy-related emissions from industrial buildings and emissions from production chemical reactions. The agriculture and forestry land sector is classified into land emissions and plant carbon sinks, while the infrastructure sector is divided by treatment object into solid waste and wastewater treatment emissions. This granular mapping, the authors contend, is what allows the model to connect spatial patterns with entity behaviors and thereby support precise planning interventions.

The system boundary for the accounting is defined consistently with IPCC guidelines, encompassing all energy-related activities and industrial process emissions occurring within the physical boundaries of the park, targeting operational-phase emissions and excluding embodied emissions from construction. Notably, emissions from on-site renewable energy production and carbon sink systems are incorporated as negative values in the final calculation. The framework also places particular emphasis on Scope 2 emissions associated with purchased electricity, heat, and steam consumed within the park, whose upstream production emissions contribute significantly to the total carbon footprint. Transport emissions are included where the park serves as origin and destination, covering employee commuting and internal logistics, while through-traffic unrelated to park activities is explicitly excluded. Emissions from waste treatment processes occurring outside the park boundaries are included, whereas upstream raw material production and transportation are excluded.

Each of the five subsystems receives its own accounting formula. The residential system combines residential building emissions with municipal treatment emissions, adjusting for vacancy rates and the energy savings achieved by low-carbon building technologies. The production system aggregates public building energy consumption across categories such as offices, hotels, and shopping malls, together with industrial manufacturing emissions estimated from annual energy consumption and fuel-specific emissions factors. The transportation system calculates commuting emissions from travel distance, mode, commuter numbers, and commuting days, and freight emissions from freight volume and distance. The renewable energy production system quantifies emissions reductions from distributed energy supply, photovoltaics, and heat pumps, while the carbon sink system estimates sequestration from ground-level, rooftop, and vertical greening, using vegetation coverage ratios, carbon sink factors for trees, shrubs, lawns, and wetlands, and a growth stage correction factor.

To make the framework operational, the researchers developed a climate-adaptive and factor-adjustable accounting module organized into four functional layers: hardware, data, computation, and rendering. The module constructs a differentiated parameter library based on China’s five climate zones, automatically identifying the local zone via GIS coordinates and invoking corresponding parameters. For residential building energy consumption, it adopts per-unit-area electricity consumption constraints specified in the national standard GB/T 51161-2016, which for instance specifies 45 kWh per square meter per year for heating in severe cold regions versus 25 kWh in hot-summer/cold-winter zones. Waste-related emissions follow a per-capita daily generation coefficient of 0.85 to 1.2 kilograms drawn from CJ/T 106-2016, and modal split assumptions for transportation derive from national passenger transport reports or local travel surveys. Users may customize parameters using local statistical yearbooks or measured data, with the system archiving modification logs to ensure transparent, auditable computation workflows.

The validation case is the Shanghai Caohejing Development Park, established in 1984 and spanning 5.98 square kilometers, of which a 3.11 square kilometer low-carbon practice zone was selected for accounting. This zone accommodates 156,000 employees across 188 buildings totaling 3.0304 million square meters of floor area, and comprehensive carbon inventory data from 2019 to 2021 was available for cross-verification. Theoretical accounting estimated residential system emissions at 4,011 tonnes per year, production system emissions at 256,062 tonnes per year, transportation emissions at 3,470 tonnes per year, renewable energy reductions of 1,210 tonnes per year, and carbon sink absorption of 640 tonnes per year. Actual measurements based on operational data showed residential, production, and transportation emissions of 4,300, 272,296, and 3,763 tonnes respectively, renewable energy reductions of 1,200 tonnes, and carbon sink absorption of 635 tonnes.

Comparative analysis showed accuracy rates between 92 and 99 percent across the individual systems. Total simulated emissions reached 264,000 tonnes, while measured emissions were 281,000 tonnes, achieving an overall accuracy of 94 percent. Errors in the residential, production, and transportation systems were approximately 7 percent, while errors in the carbon sink and renewable energy production systems were only around 1 percent. The authors attribute the remaining discrepancies primarily to dynamic factors such as building vacancy rates and the proportion of low-carbon buildings, which are precisely the variables their framework is designed to capture through monitoring rather than assume. They conclude that the proposed methods improve the precision, accountability, and planning applicability of low-carbon industrial parks, providing a practical decision-support toolbox for low-carbon planning, construction, and governance, while noting that future work can involve continuous tracking of carbon flows across the five systems to enable finer-grained analysis and deepen understanding of the factors that drive park-scale emissions.

Subject of Research: Carbon flow modeling and emissions accounting for low-carbon industrial parks

Article Title: Research on carbon flow modeling and carbon emissions accounting methods for low-carbon industrial parks

Article References: Sun, J., & Wu, T. (2026). Research on carbon flow modeling and carbon emissions accounting methods for low-carbon industrial parks. Frontiers of Urban and Rural Planning, 4(1), Article 8. https://doi.org/10.1007/s44243-026-00080-x

Image Credits: AI Generated

DOI: 10.1007/s44243-026-00080-x

Keywords: carbon flow, industrial parks, carbon emissions accounting, system dynamics, low-carbon planning, Shanghai Caohejing, carbon sinks, renewable energy, Scope 2 emissions, IPCC guidelines, urban planning, carbon neutrality