Li and Yan (2024, Land) analyzed the central urban area of Changxing County (China) for the year 2018 and found that block-scale density factors — above all floor area ratio (FAR), interacting non-linearly with building density, commercial outlet density, and population density — explain most of the spatial variation of the carbon emission intensity of construction land, and derived suitable density intervals as planning regulation targets. The authors explicitly left open whether these relationships are stable over time as density, technology, energy mix, policy, and the economy evolve, whether they generalize beyond one county-level city in the Yangtze River Delta to other small cities in different areas, and whether they survive complete building-operational energy accounting that includes non-electric uses; the further question of whether they extend to larger cities, other climates, and other building stocks follows from the same stated generalizability limitation. This problem asks for a determination of that external validity: whether, how, and under what conditions the identified factor hierarchy, dominant interactions, association chains, and suitable density intervals remain stable or change across time, urban contexts, and accounting scope.

Background

The carbon emission intensity of construction land (CICL) is the amount of carbon dioxide emitted per unit area of urban developed land, in kgCO2 per square meter. A long line of research relates CICL to urban density — how concentrated buildings, people, and activities are — but most studies work at the scale of kilometer grids or whole cities, treat density factors one at a time, and derive emissions indirectly by allocating city-level statistics to pixels. Blocks (the individual land parcels bounded by streets, carrying one principal function such as residential, commercial, or public) are the units at which urban development is actually planned and regulated, so block-scale evidence is directly usable for planning, yet it has been scarce because measured block-level energy data are hard to obtain.

The geographical detector is a spatial-statistical method built for spatial stratified heterogeneity: it stratifies the study area by each candidate explanatory factor and computes a qq-statistic between 0 and 1 measuring the share of the variance of the response variable that the stratification explains; an interaction module compares the qq of two factors individually with the qq of their joint stratification, classifying the pair as, for example, bi-variate enhanced or non-linearly enhanced (joint explanatory power exceeding the sum of the individual powers). The optimal parameters-based geographical detector (OPGD) is a refinement (Song et al. 2020) that searches, for each factor, over classification methods and numbers of class intervals (e.g. 3 to 10) to maximize qq, removing arbitrary discretization choices; its risk-detection module then reports which strata (intervals) of each factor carry low response values, yielding "suitable ranges" that planners can use as regulation targets.

Li and Yan (2024) applied OPGD to the central urban area of Changxing County for 2018, using measured annual electricity consumption of sampled blocks (electricity accounted for over 93% of sampled block operational energy, with most sampled blocks above 99%) converted to carbon emissions with the local grid emission factor of 0.8463 kgCO2/kWh. They built a density system of physical-environmental factors (floor area ratio (FAR), i.e. built floor area divided by plot area; building density (BD), the ground coverage of buildings; average and standard deviation of building height; building form and quantity descriptors; road network density; green rate) and socioeconomic factors (population density (PD), commercial outlet density (COD), building age), and analyzed residential (RB), commercial (CB), and public (PB) blocks separately. Findings: FAR was the single most influential factor everywhere (qq-values 0.7086, 0.5378, 0.5652 for RB, CB, PB); the dominant interactions were FAR with BD for residential blocks (q=0.8694q=0.8694), FAR with COD for commercial blocks (q=0.9127q=0.9127), and FAR with PD for public blocks (q=0.9396q=0.9396), all non-linear enhancements; several factors (e.g. BD, COD) had no individual effect but strong interactive effects; and per-block-type "association chains" of interacting density factors — ordered chains constructed, for each block type, by ranking pairwise interaction explanatory power and applying the OPGD risk-detection module to the same sampled blocks — were extracted as the basis for collaborative planning control, with suitable density intervals from risk detection serving as numeric regulation targets.

The authors state three limitations that jointly define an external-validity question. First, the analysis is a single-year (2018) cross-section, so nothing is known about how the relationships shift as density, technology, energy mix, policy, and the economy change over time. Second, the sample is one county-level city in the Yangtze River Delta, so whether the factor importances, interaction strengths, association chains, and interval targets transfer to other small cities in different areas — and, more broadly, to larger cities and other climates, economic structures, and building stocks — is unresolved. Third, electricity consumption proxies total operational energy, so the effect of omitting non-electric uses (direct fossil fuel use, district heating) and of measurement error on the estimated intensities and detected factor structure is unknown. Later reviews treat these as field-wide gaps rather than settled results: a 2026 systematic review of over one hundred urban-morphology-and-carbon studies documents that morphological thresholds such as optimal FAR and compactness ranges vary with city size, climate zone, and urbanization stage and expressly cautions against transferring them between contexts, while noting that the comparable cross-scale, cross-context studies needed to establish whether and how a particular city’s structure transfers remain scarce. Whether the specific Changxing 2018 baseline structure transfers to other contexts, and under what conditions, is therefore unresolved, and comparable multi-year block-level and non-electric end-use data remain scarce.

Problem Statement

Determine the external validity of the block-scale density–carbon-emission-intensity relationships identified by Li & Yan (2024) for Changxing County in 2018 — the hierarchy of factor explanatory powers, the dominant pairwise interactions (FAR with building density for residential blocks, FAR with commercial outlet density for commercial blocks, FAR with population density for public blocks), the extracted association chains, and the suitable density intervals used as planning regulation targets. Concretely, with the Changxing 2018 results as the explicit baseline: (i) establish how these relationships and intervals change over time at the block scale as density, technology, energy mixes, policy, and economic conditions evolve; (ii) establish whether and how they differ across urban contexts — larger cities, other small cities, and regions with different climates, economic structures, and building stocks — at a comparable block scale; and (iii) quantify how the estimated intensities and the detected factor structure change when the electricity proxy is replaced by complete building-operational energy accounting that includes non-electric uses (direct fossil fuel use, district heating) and accounts for measurement error. The requested outcome is a quantified determination, for each of the three axes, of which elements of the baseline structure (factor importances, interaction types and strengths, interval targets) are stable and which are context- or time-dependent, together with the conditions under which density-regulation intervals derived in one context can be transferred to spatial planning in another. The urban contexts named under (ii) delimit the scope of the cross-context comparison rather than requiring a separate study for every named context.

References

  1. Liutong Li, Fengying Yan. How Does Density Impact Carbon Emission Intensity: Insights from the Block Scale and an Optimal Parameters-Based Geographical Detector. Land 2024, 13(7), 1036. DOI: 10.3390/land13071036. https://doi.org/10.3390/land13071036
  2. Yongze Song, Jinfeng Wang, Yong Ge, Chengdong Xu. An optimal parameters-based geographical detector model enhances geographic characteristics of explanatory variables for spatial heterogeneity analysis: cases with different types of spatial data. GIScience & Remote Sensing 2020, 57(5), 593–610. DOI: 10.1080/15481603.2020.1760434. https://doi.org/10.1080/15481603.2020.1760434
  3. Leshui Huang, Linxuan Xie, Meng Cai. The Impact of Urban Morphology on Carbon Emissions Under Urban Renewal: A Critical Review. Land 2026, 15(6), 1033. DOI: 10.3390/land15061033. https://doi.org/10.3390/land15061033
  4. Bingxin Li, Qiang Zheng, Xue Jiang, Chennan He. Multi-Scale Spatial Structure Impacts on Carbon Emission in Cold Region: Case Study in Changchun, China. Sustainability 2024, 17(1), 228. DOI: 10.3390/su17010228. https://doi.org/10.3390/su17010228
  5. Yong Yu, Sameer Kumar, Zhu Ye. Exploring the interaction and driving factors of urban compactness and carbon emission intensity in the Pearl River Delta Urban Agglomeration. Scientific Reports 2025, 15, 10091. DOI: 10.1038/s41598-025-94435-x. https://doi.org/10.1038/s41598-025-94435-x
  6. Yuchen Qin, Yuchen Xie, Yichen Dong, Qi Zhang, Xiaoping Li, Qiwei Chen, Minghao Wang, Shen Xu. The synergistic effects of multiple urban morphological factors on building and residential travel carbon emissions at the neighborhood scale. Energy and Buildings 2026, 361, 117414. DOI: 10.1016/j.enbuild.2026.117414. https://doi.org/10.1016/j.enbuild.2026.117414
  7. Wang Liu, Zhipeng Zhu, Zhongli Lin, Yingxue Wen, Xiong Yao. Fine-scale delineation of urban carbon emission zones using SDGSAT-1 glimmer imagery and block-scale morphology. Energy and Buildings 2026, 361, 117441. DOI: 10.1016/j.enbuild.2026.117441. https://doi.org/10.1016/j.enbuild.2026.117441
  8. Ruijun Chen, Chen Ren, Junqi Wang, Shi-Jie Cao. From buildings to regions: a cross-scale review of spatial morphology and its dynamic relationship with carbon emissions. Energy and Buildings 2026, 363, 117586. DOI: 10.1016/j.enbuild.2026.117586. https://doi.org/10.1016/j.enbuild.2026.117586