The study, published in Frontiers in Environmental Science, distinguishes between what researchers call observed peaks and effective peaks. An observed peak is simply the highest recorded energy consumption within the study period. An effective peak, by contrast, requires that the city’s average energy consumption over the final three years be at least 5 percent below its historical maximum. This distinction matters because a city that records its highest consumption in 2022—the final year of the sample—cannot be confirmed as having peaked. Its observation is right-censored rather than confirmed, a statistical problem that affects 202 of the 336 cities, or 60.12 percent of the sample.

The gap between observed and effective peaks is striking. While 202 cities recorded their historical maximum in 2022, only 97 cities had an observed peak no later than 2019. Applying the stricter effective-peak rule reduces the count further to 94 cities, or 27.98 percent of the sample. The researchers then classified cities into four transition stages: stable post-peak, rebound-risk, still-rising or near-peak expansion, and fluctuating transition. Only 88 cities (26.19 percent) are classified as stable post-peak, while 208 cities (61.90 percent) remain in the still-rising or near-peak expansion category. Seven cities are classified as rebound-risk cases—cities that peaked earlier but whose 2022 energy consumption had returned close to the historical maximum, with a median rebound index of 0.943, meaning approximately 94 percent of the decline from peak to post-peak minimum had been recovered by 2022.

The regional divide is equally pronounced. In eastern China, 46 of 87 cities satisfy the effective post-peak criterion, a share of 52.87 percent. By comparison, the corresponding figures are 25.61 percent for central China, 16.03 percent for western China, and 16.67 percent for the northeast. The right-censoring problem is especially severe in the west and northeast: 107 of 131 western cities (81.68 percent) and 29 of 36 northeastern cities (80.56 percent) recorded their historical maximum in 2022. Eastern cities were, on average, 21.61 percent below their historical peak in 2022, while the equivalent figures in central, western, and northeastern China were 4.99 percent, 3.12 percent, and 3.38 percent, respectively. Formal statistical tests confirm these differences are systematic, with a chi-square test rejecting equality in effective peak status across regions (χ² = 38.56, p < 0.001).

The study’s methodology is notable for its use of satellite nighttime-light data to reconstruct city-level energy consumption. Because consistent city-level energy statistics are not available for the full sample period, the researchers combined nighttime-light observations from the Defense Meteorological Satellite Program’s Operational Linescan System (DMSP-OLS) and the Visible Infrared Imaging Radiometer Suite (VIIRS) with officially reported provincial energy statistics. The approach matches provincial nighttime-light intensity with reported provincial energy consumption to estimate the relationship between light and energy use, then applies this relationship to city-level light totals. The resulting estimates are expressed in standard coal equivalent and capture total city-level energy consumption without distinguishing energy use by sector or fuel type.

The researchers conducted extensive robustness checks to ensure their findings are not driven by arbitrary parameter choices. Varying the tolerance threshold from 5 percent to 0, 3, or 10 percent changes the count of effective post-peak cities only modestly, from 94 to between 75 and 97. Extending the persistence window from three to four years lowers the count to 71. Applying a three-year moving average to reduce sensitivity to annual fluctuations increases the number of cities peaking in 2022 from 202 to 212 while reducing effective post-peak cities from 94 to 76. Across all alternative specifications that retain 2022 as the sample endpoint, the share of effective post-peak cities remains below one-third of the sample.

A pandemic-period stress test reveals how sensitive peak classifications are to the sample endpoint. Using 2019 as the endpoint, 155 cities satisfy the effective-peak criterion—substantially more than the 94 identified in the full-sample classification. The researchers then tracked these 155 pre-pandemic effective-peak cases through 2020–2022 using each city’s pre-2020 historical maximum as the benchmark. Only 60 of the 155 cities remained below that benchmark in all three years, whereas 95 exceeded it at least once. By 2022, 94 of these 155 pre-pandemic cases had returned to within 5 percent of their pre-pandemic historical maximum or had moved above it. The researchers treat this period as a stress test of peak durability rather than evidence of a causal pandemic effect.

Trend-break analysis provides complementary evidence about the nature of urban energy transitions. The researchers fitted a single-break trend model to the logarithm of energy consumption for each city, classifying cities as exhibiting acceleration, slowdown-but-still-growing, or reversal. The results show that 246 cities (73.21 percent) fall into the slowdown-but-still-growing category, while only 73 cities show a reversal pattern and 17 exhibit acceleration. This means the dominant urban energy-transition pattern in China is growth deceleration rather than absolute decline. Many cities are growing more slowly than before, but they have not yet entered a stable post-peak stage—a distinction with substantial policy implications, since cities with slower growth still need infrastructure, clean energy supply, and demand management, whereas stable post-peak cities need measures to consolidate decline and prevent rebound.

The distinction between energy peaks and carbon peaks is central to the study’s contribution. A city’s emissions can decline because of cleaner electricity, fuel substitution, industrial upgrading, or imported power—even while total energy use continues to rise. Conversely, a drop in absolute energy consumption may point to deeper structural shifts: reduced reliance on heavy industry, improved efficiency, more compact urban form, or demand saturation. The researchers argue that city-level carbon-emission peaking and energy-consumption peaking should be treated as complementary rather than interchangeable dimensions of urban transition. For energy security, infrastructure planning, and demand-side management, the absolute energy-consumption peak is a separate and policy-relevant object of inquiry.

The findings arrive at a critical moment for China’s climate commitments. The country’s updated nationally determined contribution commits to peaking carbon emissions before 2030 and achieving carbon neutrality before 2060, while the 14th Five-Year Plan targets a 13.5 percent reduction in energy intensity and an 18 percent cut in carbon intensity between 2021 and 2025. The researchers emphasize that their framework is diagnostic rather than causal—it identifies where cities stand in the energy transition without estimating the effects of particular economic or policy drivers. They suggest that the transition categories can guide demand-scenario selection in planning exercises: for stable post-peak cities, the recent post-peak demand level can serve as the central planning case; for rebound-risk cities, the historical peak should remain within the planning range; and for still-rising cities, the central scenario should not assume an imminent absolute decline. The study also notes that the four-region classification captures broad spatial gradients but cannot substitute for a functional classification based on industrial or resource characteristics, since cities at different transition stages coexist within each region.

Subject of Research: Urban energy peaking in China: timing, persistence, rebound risk, and regional heterogeneity

Article Title: Urban energy peaking in China: timing, persistence, rebound risk, and regional heterogeneity

Article References: Liu, Y., Lau, N. W., Lam, J. F. I., & Chen, G. (2026). Urban energy peaking in China: timing, persistence, rebound risk, and regional heterogeneity. Frontiers in Environmental Science, 14, Article 1963897. https://doi.org/10.3389/fenvs.2026.1963897

Image Credits: AI Generated

DOI: 10.3389/fenvs.2026.1963897

Keywords: China, energy transition, urban energy, carbon peak, nighttime lights, regional heterogeneity, rebound risk, climate policy, Urban, energy, peaking, timing

A research team led by Huijun Wu and Chang Su constructed what they call the IEC nexus, a unified framework linking industrial structure, ecological environment and carbon emissions into a single analytical system. The approach borrows from the ‘nexus’ thinking first proposed at the 2011 Bonn Conference for water, energy and food, in which interdependent systems mutually reinforce or constrain one another. Rather than treating economic transformation, pollution and greenhouse gases as separate policy files, the team measured all three simultaneously for Huainan between 2000 and 2022, using entropy-weighted indicators drawn from national and provincial statistical yearbooks, energy inventories and local environmental bulletins. Carbon emissions were calculated following IPCC guidelines, converting the physical consumption of coal, oil and natural gas into carbon dioxide using fuel-specific heating values, carbon contents and oxidation rates.

The centerpiece of the analysis is the coupling coordination degree, a metric that ranges from zero to one and captures how well the three subsystems evolve together rather than at each other’s expense. The results tell a story in three acts. In the early period from 2000 to 2008, the industrial subsystem lagged badly behind, with its evaluation index fluctuating between just 0.13 and 0.44 while environmental and carbon indices ran higher. During a difficult transition from 2008 to 2016, the accumulated costs of coal dependence became visible: the coordination degree actually slipped from 0.753 in 2012 to 0.741 in 2016 as environmental quality briefly deteriorated. Then, from 2016 onward, the system entered what the authors describe as a high-quality development stage, with the overall coordination degree climbing from 0.552 in 2000 to 0.832 in 2022, crossing from ‘barely coordinated’ into the ‘good coordinated development’ band.

Perhaps the most striking finding is how lopsided the progress has been. The industrial structure index surged to about 0.52 by 2020 and the ecological environment index climbed to roughly 0.94, yet the carbon emissions index crawled along at only about 0.40. In other words, Huainan has learned to restructure its economy and clean its air far faster than it has learned to decarbonize. The coupling degree itself remained stubbornly high throughout, ranging from 0.754 to 0.991, which the researchers interpret as evidence that the three subsystems are locked into tight mutual dependence: improvements in one cannot be sustained for long without corresponding movement in the others.

To identify what was actually holding the city back, the team applied an obstacle degree model that decomposes the coordination deficit into individual indicators. The verdict shifted dramatically over the two decades studied. In 2000, the industrial structure subsystem accounted for a dominant 71.04 percent of the total obstacle, driven largely by weak tertiary-sector employment and stagnant rural incomes. By 2022 that share had fallen to 45.07 percent, while the ecological environment subsystem more than doubled its contribution from 15.03 percent to 37.48 percent, and carbon emissions rose from 13.93 percent to 17.46 percent. The bottleneck has migrated: as structural constraints eased, pollution intensity, particularly sulfur dioxide emissions per unit of GDP, and the lingering weight of coal-linked industry became the new front line.

The study then turned from diagnosis to prophecy, pairing the STIRPAT model, a regression framework relating environmental pressure to population, affluence and technology, with a Grey GM(1,1) forecasting model. Because the six explanatory variables, including population, per capita GDP, urbanization, secondary-industry share, energy intensity and energy structure, showed substantial multicollinearity, the team calibrated the model with ridge regression, setting the ridge parameter at 0.14 and achieving an R-squared of 0.868 with an average fitting error of 6.18 percent. The authors are careful to note that only the population coefficient reached even marginal statistical significance, so the projections should be read as conditional scenarios rather than firm causal predictions.

Those scenarios, however, are revealing. Under a business-as-usual pathway, Huainan’s carbon emissions would not peak until around 2035, five years past the national target, at roughly 29.6 million tons, about 1.27 times the 2020 level. Merely optimizing industrial structure barely improves that timing. The low-carbon and energy-saving scenario, which combines faster cuts in energy intensity with a cleaner energy mix, brings the peak forward to 2030 at about 28.7 million tons while preserving steady economic growth, making it the pathway the authors judge most realistic and aligned with China’s dual carbon goals. An aggressive green development scenario peaks earliest, in 2028 at just 24.9 million tons, but only by deliberately throttling population growth and urbanization. At the other extreme, an extensive development scenario that prioritizes raw economic expansion delays the peak to 2040 at a hefty 35.1 million tons.

The Grey model projections for the coordination degree itself add a further layer of nuance. Across all scenarios the index continues rising through 2036, but the green development pathway approaches perfect coordination by 2032, followed by industrial structure optimization, then the low-carbon energy pathway, with the extensive scenario trailing the field. The ordering suggests that structural adjustment delivers the quickest coordination gains, while energy-focused measures, though slower to register, are what actually bend the emissions curve. For a city where coal mining has caused land subsidence, water pollution and ecosystem degradation across some of eastern China’s most representative subsidence zones, both fronts matter.

The policy implications the authors draw are refreshingly specific. Huainan should not simply shrink its secondary sector, they argue, but upgrade its internal composition, cultivating new energy equipment manufacturing, photovoltaic and wind industries, digital economy sectors and modern logistics, while converting mining subsidence zones into wetland parks and ecological conservation areas. Coal-power plants should pursue ultra-low emission retrofits and carbon capture technologies, and the city should build a dynamic monitoring system around the very indicators the obstacle analysis flagged: tertiary employment, pollution intensity, energy intensity and energy structure. Because the constraints themselves shift over time, the authors contend, governance must be adaptive and coordinated across industrial, environmental, energy and planning departments rather than siloed.

The study is candid about its limits. A 2016 administrative boundary change, when Shou County was incorporated into Huainan, complicates comparisons of population-related indicators across the time series, and some missing data were filled by regression-based imputation. The findings from a single coal city may not transfer cleanly to regions with different resource endowments, and the forecasting models rest on assumptions about parameter stability that real-world shocks could upend. The authors also stress that their correlations describe temporal associations, not proven causal effects. Even so, the broader message travels well beyond Anhui: for the hundreds of resource-dependent cities worldwide staring down decarbonization deadlines, Huainan’s two-decade ledger suggests that industrial restructuring, ecological restoration and carbon mitigation can indeed rise together, but only if energy intensity and energy structure are treated as the decisive battleground rather than an afterthought.

Subject of Research: Coupling coordination of industrial structure, ecological environment and carbon emissions in the coal-resource city of Huainan, China

Article Title: Can coal-resource cities achieve high-quality coordination of industry, environment, and carbon? Evidence from Huainan, China

Article References: Wu, H., Su, C., Zhang, Y., Li, M., Kong, X., Liu, Y., & Gao, L. (2026). Can coal-resource cities achieve high-quality coordination of industry, environment, and carbon? Evidence from Huainan, China. Environmental and Sustainability Indicators, 32, Article 101541. https://doi.org/10.1016/j.indic.2026.101541

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101541

Keywords: Huainan, coal-resource cities, carbon emissions, coupling coordination degree, STIRPAT model, energy transition, industrial structure, ecological environment, carbon peak, obstacle degree analysis, China dual carbon goals, GM(1,1) forecasting