Environmental regulations are increasingly fine-grained and spatially differentiated, but the resulting text-intensive and largely unstructured policy documents make regulatory measures difficult to interpret consistently. Here, a regulatory intensity evaluation framework based on Large Language Models (LLMs), combining policy authority, target stringency, and measure clarity was developed to quantify China's Ecological and Environmental Zoning-based Regulation (EZR). Over 20 million Chinese characters from lists of environmental permits (LEPs) were operationalized for 39,928 integrated environmental zoning units. Benchmarking across models and prompting strategies indicates that DeepSeek with few-shot prompting achieves the best extraction performance, improving accuracy, precision and recall by about 14%, 20%, and 14%, respectively relative to the best zero-shot setting. The average regulatory intensity across all IEEUs is 4.16, with pronounced heterogeneity across control types and dimensions: critical regulatory units (5.10) and the spatial layout constraint dimension (7.26) exhibit the strongest intensity. A clear east–west gradient is observed, with Shanghai (6.70) and Beijing (6.26) at the upper end, while provincial-level scores extended down to 1.68. Response surface methodology (RSM) further shows that regulatory intensity reflects a joint configuration of regional development and environmental pressure, featuring nonlinear and interaction effects. This study demonstrates a scalable LLM-enabled approach for extracting large, unstructured zoning-policy texts into comparable intensity metrics, providing methodological guidance for quantitative evaluation of differentiated environmental regulation and actionable evidence for policy design and management.

Individual prevention behaviors are a primary line of defense during the early stages of novel infectious disease outbreaks, yet their adoption is heterogeneous and difficult to forecast-especially when empirical data are scarce and epidemic-policy contexts evolve rapidly. To address this gap, we develop an LLM-based prevention-behavior simulation framework that couples (i) a static module for behavior-intensity prediction under a specified external context and (ii) a dynamic module that updates residents' perceived risk over time and propagates these updates into behavior evolution. The model is implemented via structured prompt engineering in a first-person perspective and is evaluated against two rounds of survey data from Beijing residents (R1: December 2020; R2: August 2021) under progressively realistic data-availability settings: zero-shot, few-shot, and cross-context transfer. Using Kolmogorov-Smirnov tests to compare simulated and observed behavior distributions (p > 0.001 as the validity criterion), the framework demonstrates robust performance and improves with limited reference examples; reported predictive accuracy increases from 72.7

Clarifying the characteristics of industrial spatial agglomeration within urban agglomerations and the associated impact mechanism on carbon emission intensity is crucial for formulating regional collaborative carbon reduction strategies. Focusing on China’s 18 major urban agglomerations and employing methods such as the location quotient (LQ) index and the spatial Durbin model (SDM), this study analyzes the features of industrial spatial agglomeration and the influence of that agglomeration on carbon emission intensity, from 2002 to 2023. The key findings are: (1) Industrial spatial agglomeration in Chinese urban agglomerations is pronounced, with the pattern of spatial agglomeration gradually shifting from single-point concentration to clustered agglomeration. (2) While total carbon emissions (TCE) and per capita carbon emissions (PCCE) in these regions increased significantly, carbon emission intensity per unit output exhibited a continuous declining trend. Regions with higher carbon emission intensity showed a tendency to shift towards urban agglomerations in northern China. (3) An analysis of core explanatory variables reveals the multifaceted impact of industrial agglomeration on carbon emission intensity. Specialized industrial agglomeration (SIA) reduces emission intensity through technology spillovers, but SIA is accompanied by rebound effects and carbon transfer to neighboring areas. Technology-intensive agglomeration (TIA) increases local emission intensity while decreasing TCE and PCCE. However, TIA also faces challenges in reducing emission intensity and issues of spatial spillover. An analysis of control variables identifies technological innovation as the core driver for reducing carbon emissions. (4) Based on these findings, carbon reduction strategies tailored to the industrial spatial agglomeration processes in urban agglomerations are proposed.

Encouraging pro-environmental behavior remains a major challenge for sustainable cities. Conventional feedback nudges can show individuals how their current behavior compares with environmental goals but often provide limited guidance on what to do differently in daily life. This study examines whether supplementing weekly feedback on participants' behavior with LLM-generated personalized action suggestions improves pro-environmental behavior, using daily electricity and hot-water conservation as a case study. We developed an LLM agent that generated weekly conservation messages from participant profiles, recent consumption records, and prior interaction history, combining a usage report with personalized suggestions, behavioral-change scenarios, and estimated savings. The agent was evaluated in a three-arm randomized field experiment with 233 university residents in Beijing from November 2024 to January 2025. Participants received text-based nudges, image-enhanced nudges, or LLM-generated personalized nudges over five intervention rounds. Daily electricity use and shower hot-water use were measured using dormitory meter readings and billing records. Compared with text-based feedback, LLM-generated personalized nudges reduced electricity consumption by 0.56 kWh per room-day (p = 0.014), corresponding to an 18.3 percentage-point higher saving rate. Image-enhanced feedback alone showed no clear improvement. Hot-water savings followed the same direction but were smaller and less precisely estimated (9.8 percentage points, p = 0.087). Personalized nudges contained more planning, appliance-specific, and action-oriented language and were associated with more sustained, task-focused engagement. These findings offer a pathway for integrating generative AI into sustainable urban management.

The cooling structure of the air-cooled turbine first-stage inlet guide vane is complex, requiring a refined multi-factor design. This study focuses on the impingement insert in the forward cavity of turbine vanes. A multi-objective optimization method is developed to improve the internal cooling structure, considering the coupled effects of jet hole diameter, streamwise spacing, and spanwise spacing. Latin Hypercube Sampling (LHS) method is employed for the design of experiments, and a Backpropagation Neural Network (BPNN) is developed to model the relationship between the geometric parameters of the impingement insert and the area-averaged Nusselt number ( Nu ), uniformity index (UI), and overall pressure drop (ΔP). Furthermore, multi-objective optimization of Nu , UI and ΔP is performed using the NSGA-II algorithm. The results show that the Pareto-optimal solutions achieve an approximately 20 Nu and 5.46

Aquatic ecological deterioration challenges watershed governance worldwide, requiring heightened attention. Local administrators' policy salience -captured by Governmental Attention (GA) derived from official documents -could shift with superior-level agenda signals and/or ecological quality. These two drivers are interpreted as politically driven orientation and problem-driven orientation in Multiple Streams Framework (MSF). However, the relative contributions of these orientations to GA remain unclear. This study quantifies their contributions to aquatic ecology governance by analyzing 3110 policies from 2020-2024 in China's Nine-Plateau-Lake (NPL) region using text mining. Results show that GA typically peaks in the 4th quarter (October-December), aligning with administrative evaluation cycles rather than natural seasons. The region exhibits more intensive yet less stable GA in 2023-2024, with the average GA doubling from 0.1 to 0.2 (maximum 1) while Coefficient of Variance (CV) increasing from 1.5 to 1.9. Considering time-lagged effects, superior agenda signals proxied by provincial-level GA (PGA) presents significant influences on GA at monthly scale, and in three out of five cities. PGA exerts higher impacts on GA in more cities than problem-driven orientation (proxied by the algae blooming area proportion), and with higher influence coefficients |beta| >= 0.12 v.s. |beta| <= 0.02. The prevalence of politically driven orientation over problem-driven orientation implies challenges in achieving the "problem-oriented governance" in aquatic ecology governance as promoted. This study develops a multidimensional, text-derived GA measure and extends MSF to quantitative analysis in watershed governance, demonstrating its potential in revealing governmental decision-making orientations and informing policy design for aquatic ecology-centered watershed governance.

The ALPS treated water has been discharged into the Pacific Ocean since August 2023. This study investigates this discharge using a newly developed three-dimensional dispersion model that incorporates migration, diffusion, and decay processes of radionuclides. A simulation over ten years is conducted using reanalyzed oceanographic data. The results indicate that tritium released from Fukushima primarily disperses eastward along the 35 degrees N latitude line. In later stages, local concentration peaks emerge in the northeastern Pacific, exceeding those in the northwest Pacific. For the vertical distribution, the tritium is generally reduced greatly with depth, but displays maximum values at subsurface layer (similar to 50m) in some regions. The concentration reaches a steady state over time, defined as the characteristic concentration, whose spatial distribution and attainment time are detailed. For major fishing grounds in the Pacific Ocean, the Hokkaido area shows the highest tritium levels, followed by Hawaii, California, Zhoushan, the Korean Peninsula, Mexico, the Philippines, Alaska, and Peru in descending order. Critically, simulated tritium concentrations in most North Pacific regions (similar to 0.01 Bq/m(3)) remain orders of magnitude below natural background levels (similar to 50 Bq/m(3)). This research elucidates threedimensional radionuclide dispersion mechanisms in global oceans, providing a quantitative methodology for future marine emergency response and contributing to long-term marine conservation efforts.

Model evaluation is crucial for verifying model credibility, especially in decision-making. Successful environmental modelling requires not only self-proved credibility from model developers/users and peer-appraised credibility from technical experts, but also decision-maker and public confidence in model credibility. We propose a participatory model evaluation approach for environmental decisions, combining the standard evaluation procedure, data-augmented peer review and multi-stakeholder engagement. To facilitate this approach, we developed DPMODE (Decision Procedure Management of surface water mODel Evaluation), a web-based system with supporting tools and database. DPMODE evaluates surface water models and recommends credible models and customized test datasets for watershed management. A case study on the Soil and Water Assessment Tool (SWAT) for the Chishui River watershed management demonstrated the effectiveness of this approach. This participatory evaluation would be an adaptive, iterative process to improve stakeholder acceptance, enhance model-based outcomes, and foster better decision pathways.

Widely existing and spatiotemporally characterized urban neighborhood noise can affect the quality of life of residents adversely. The number of neighborhood noise complaints (NNCs) is a representative indicator of residents affected by neighborhood noise and can be influenced by various built environment factors. This study aimed to identify the influencing built environment factors and their impact variations over time and space on NNCs. A total of 194,558 NNCs in the main urban area of Beijing, China, were selected as case studies. The results indicated that NNCs had temporal fluctuations and strong spatial clustering. Besides points of interest (POI) like restaurants and life services and land use variables, the community existence year and plot ratio also had significant positive impacts on the number of NNCs. Temporally, variable impacts showed U-shaped trends hourly with peaks around 10:00 to 14:00, while the coefficients of community existence year and plot ratio exhibited night rebounds with an increase after 22:00. Spatially, variable impacts displayed significant spatial variations with a centralized clustering of high-values. These spatiotemporal disparities underscore the need to consider comprehensive effects of urban built environment and residents’ living routines/perceptions of neighborhood noise. The findings of this study provide valuable insights for the refined and integrated urban neighborhood noise control strategies.

China's Integrated Environmental Zoning (IEZ) policy, initiated in 2021, targets industrial restructuring by enforcing specific regulations for each Integrated Environmental Unit (IEU) regarding spatial arrangement, emissions, and technological efficiency, among others. However, firms, as the fundamental operational units, may react differently to IEZ interventions, potentially causing unpredictable socioeconomic side effects. This study employs an agent-based model to assess the impacts of IEZ on industrial outputs, emissions, efficiency, and spatial layout in Hebei Province. The findings reveal that (1) IEZ fosters sustainable industrial growth (12.1 %) and cuts SO2 emissions by 9.6 % by 2030 compared to the baseline scenario, primarily through the elimination of non-compliant industries and consistent technological advancements; (2) IEZ significantly suppresses air-pollution-intensive industries and considerably boosts low-emission sectors like equipment manufacturing, but has limited effects on water-pollution-intensive industries; (3) IEZ encourages firm transfers and agglomeration within IEUs, with intra-city transfers being the predominant form of industrial spatial restructuring (accounting for 90.9 %). This study suggests enhancing water pollution efficiency regulations within the IEZ and facilitating inter-city industrial transfers to optimize provincial industrial layout.

Urban water systems (UWSs) continuously evolve in response to changes in urban populations, technological advancements, and lifestyle shifts, resulting in significant changes in greenhouse gas (GHG) emissions. Understanding how GHG emissions vary across the different developmental stages of a UWS is crucial for charting pathways toward carbon neutrality under varying levels of urbanization and infrastructure maturity. To explore the long-term patterns of GHG emissions from the UWS, we developed a systematic accounting framework encompassing four energy-related subsystems: water extraction, water supply, residential water use, and wastewater treatment. We applied this framework to China’s UWS across its transitional trajectory—from early development to system-wide maturity (1980–2020) at the provincial level. Results show that over the 40 years, GHG emissions from China’s UWS increased approximately 14-fold, surpassing the overall rate of population growth by 143.9%. From the early 1990s till now, residential water use emerged as the dominant source of UWS-related emissions, accounting for approximately 77.6% of total emissions. Our scenario analysis estimates a potential 34.0% reduction in China’s carbon emissions (128.3 Mt CO2-eq) by 2030 through water-saving strategies. This study offers critical insights into promoting low-carbon operations and sustainable management of UWS, and serves as an important reference for global efforts net-zero water infrastructure.

The increasing amount of pressure related to water and energy shortages has increased the urgency of cultivating individual conservation behaviors. While the concept of nudging, i.e., providing usage-based feedback, has shown promise in encouraging conservation behaviors, its efficacy is often constrained by the lack of targeted and actionable content. This study investigates the impact of the use of large language models (LLMs) to provide tailored conservation suggestions for conservation intentions and their rationale. Through a survey experiment with 1,515 university participants, we compare three virtual nudging scenarios: no nudging, traditional nudging with usage statistics, and LLM-powered nudging with usage statistics and personalized conservation suggestions. The results of statistical analyses and causal forest modeling reveal that nudging led to an increase in conservation intentions among 86.9 increase of 18.0 88.6 to LLM-powered nudges enhances self-efficacy and outcome expectations while diminishing dependence on social norms, thereby increasing intrinsic motivation to conserve. These findings highlight the transformative potential of LLMs in promoting individual water and energy conservation, representing a new frontier in the design of sustainable behavioral interventions and resource management.

Classifying household water-consumption behaviors is crucial for providing targeted suggestions for watersaving behaviors and enabling effective resource management and conservation. Although it is common knowledge that energy consumption is closely coupled with household water consumption, the effectiveness of energy consumption information in classifying household water-consumption behaviors remains unexplored. This study proposes a hybrid model of long short-term memory (LSTM) and random forest (RF) using water and electricity consumption as inputs to classify household water-consumption behaviors. Data from three households in Beijing collected from January to March 2020 were used for the case studies. The hybrid model achieved a macro F1 score of 0.89 at a 5-min resolution, outperforming the standalone LSTM and RF models. Additionally, the inclusivity of time-series electricity consumption improves the accuracy (F1 scores) of classifying bathing and laundry behaviors by 0.12 and 0.20, respectively. These findings underscore the scientific value of integrating electricity consumption as a proxy variable in water-consumption behavior classification models, demonstrating its potential to enhance accuracy while simplifying data acquisition processes. This study establishes a framework for demand-side water management aimed at empowering residents to understand their own water-energy consumption behavior patterns and engage in personalized water conservation efforts.

Blending ammonia in combustion is an effective approach to reducing carbon emission for diesel engines, but the combustion of ammonia may result in the generation of elevated concentrations of nitrous oxide (N2O), potentially contributing to a new source of greenhouse gas emissions. This study investigates the impact of engine load, diesel double injection strategy, and the ammonia energy ratio (AER) on N2O emission in ammonia-diesel dual-fuel combustion mode. The results show that N2O emission is increased with blending ammonia compared to pure diesel combustion, while as the load increases, the higher in-cylinder combustion temperature leads to a decrease in N2O emission. In the ammonia-diesel dual-fuel combustion mode, diesel double injection strategy is conducive to the reduction in N2O emission compared to single injection strategy. As the pre-injection timing is delayed, N2O emission first decreases and then increases, reaching a minimum near −40 °CA ATDC. The postponed main-injection timing leads to the decreased in-cylinder combustion temperature, deteriorating N2O emission. In the engine with a compression ratio (CR) of 18, when AER is in the range of 20% to 65%, the volume fraction of N2O emission remains around 20 × 10−6. However, under the condition of AER = 80%, low chemical reactivity of ammonia causes the substantially increased N2O emission. By increasing CR to 21 and adopting optimal injection strategy, combustion in activated thermal atmosphere is achieved, resulting in a substantial reduction in the N2O original emission concentration. Ultra-low N2O emission (exhaust N2O volume fraction less than 10 × 10−6) is achieved at AER = 80%.

How to address public health priorities after COVID-19 is becoming a critical task. To this end, we conducted wastewater surveillance for six leading pathogens, namely, SARS-CoV-2, norovirus, rotavirus, influenza A virus (IAV), enteroviruses and respiratory syncytial virus (RSV), in Nanchang city from January to April 2023. Metaviromic sequencing was conducted at the 1st, 4th, 7th, 9th, 12th and 14th weeks to reveal the dynamics of viral pathogens that were not covered by qPCR. Amplicon sequencing of the conserved region of norovirus GI and GII and the rotavirus and region encoding nonstructural protein of RSV was also conducted weekly. The results showed that after a rapid decrease in SARS-CoV-2 sewage concentrations occurred in January 2023, surges of norovirus, rotavirus, IAV and RSV started at the 6th, 7th, 8th and 11th weeks, respectively. The dynamics of the sewage concentrations of norovirus, rotavirus, IAV and RSV were consistent with the off-season resurgence of the above infectious diseases. Notably, peak sewage concentrations of norovirus GI, GII, rotavirus, IAV and RSV were found at the 6th, 3rd, 7th, 7th and 8th weeks, respectively. Astroviruses also resurge after the 7th week, as revealed by metaviromic data, suggesting that wastewater surveillance together with metaviromic data provides an essential early warning tool for revealing patterns of infectious disease resurgence.

Policy making is a highly interactive process with environmental systems. Evaluation of policy's potential environmental impacts has been regarded as an efficient precautionary measure to avoid unexpected losses and environmental risks. One of the greatest challenges is how to streamline the policy-making process through an integrated evaluation approach. This paper formulated a process-based framework for environmental impact assessment of policy making (PB-EIA), by integrating policy problem identification, environmental impact assessment, and policy cost-effectiveness evaluation. Using China's wastewater resource utilization (WWRU) policy as a case study, the institutional, technological, and economic constraints or limitations related to wastewater resource utilization were identified, and the potential environmental impacts of policy implementation on energy consumption and carbon emissions were evaluated. The expected ecological and environmental benefits of the WWRU policy were supposed to be significant, and the cost of recycled water was relatively low compared with other unconventional water resources in China. Furthermore, significant variations in the environmental impacts and policy effectiveness were found across different sectors and regions. According to the evaluation results, some recommendations for supporting policy improvements and follow-up policy implementations were proposed. The specific technical procedures and application of PB-EIA were also discussed. This study demonstrates that the process-based framework can effectively incorporate environmental considerations into policy-making processes and promote sustainable development.

Estimating household water consumption can facilitate infrastructure management and municipal planning. The relatively low explanatory power of household water consumption, although it has been extensively explored based on various techniques and assumptions regarding influencing features, has the potential to be enhanced based on the water-energy nexus concept. This study attempts to explain household water consumption by establishing estimation models, incorporating energy-related features as inputs and providing strong evidence of the need to consider the water-energy nexus to explain water consumption. Traditional statistical (OLS) and machine learning techniques (random forest and XGBoost) are employed using a sample of 1320 households in Beijing, China. The results demonstrate that the inclusion of energy-related features increases the coefficient of determination (R2) by 34.0% on average. XGBoost performs the best among the three techniques. Energy-related features exhibit higher explanatory power and importance than water-related features. These findings provide a feasible modelling basis and can help better understand the household water-energy nexus.

Accurately assessing and predicting the impacts of land use changes on ecosystem carbon stocks in the Yellow River Basin (YRB) and exploring the optimization of land use structure to increase ecosystem carbon stocks are of great practical significance for China to achieve the goal of “double carbon”. In this study, we used multi-year remote sensing data, meteorological data and statistical data to measure the ecosystem carbon stock in the YRB from 2000 to 2020 based on the InVEST model, and then simulated and measured the ecosystem carbon stock under four different land use scenarios coupled with the FLUS model in 2030. The results show that, from 2000 to 2020, urban expansion in the YRB continued, but woodland and grassland grew more slowly. Carbon stock showed an increasing trend during the first 20 years, with an overall increase of 7.2 megatons, or 0.23%. Simulating the four land use scenarios in 2030, carbon stock will decrease the most under the cropland protection scenario, with a decrease of 17.7 megatons compared with 2020. However, carbon stock increases the most under the ecological protection scenario, with a maximum increase of 9.1 megatons. Furthermore, distinct trends in carbon storage were observed across different regions, with significant increases in the upstream under the natural development scenario, in the midstream under the ecological protection scenario and in the downstream under the cropland protection scenario. We suggest that the upstream should maintain the existing development mode, with ecological protection prioritized in the middle reaches and farmland protection prioritized in the lower reaches. This study provides a scientific basis for the carbon balance, land use structure adjustment and land management decision-making in the YRB.

The Water-Energy Nexus (WEN) provides a comprehensive concept for the cooperative management of resources. Although the WEN system in cities is intricately connected to socioeconomic activities, relationship between WEN and economic systems remains understudied. This study introduces a tri-dimensional Nexus Pressure Index (NPI) to assess the pressure on WEN system. Gross Domestic Product (GDP) per capita and city tiers in the urban agglomeration were used to assess the relationship between the characteristics of WEN and economic system. We conducted a case study of 296 cities in China and 1330 counties in the United States from 2012 to 2019. During the 9 year study period, on average, pressure on WEN system have relieved by 22% in China and 27% in the United States, measured by NPI. Cities with most ideal characteristics (low pressure in all dimensions) rank merely in the middle of all eight classes, with GDP per capita 74% and 85% of the highest-GDP-per-capita class in China and the US respectively. Well-performing WEN system does not yield best economic outcomes. High water pressure correlates with better economic performance in the US, while high-energy-pressure cities had GDP per capita about 50% and 70% of the class with highest GDP per capita in China and the US, respectively, suggesting stronger economic constraints from energy stress. Urban agglomeration analysis revealed a negative relationship between WEN and economic performance. NPI in emerging cities is 0.6-1 lower than NPI in regionally-central cities in China, while 0.2-0.5 lower in the US. These results underscore the contradiction between preferred WEN characteristics and higher economic performance, and underpin the resource curse hypothesis at city-level in the two considered giants. A sustainable approach to harmonize WEN and economic system is in urgent need.