International Journal of Global Warming (32 papers in press)

Regular Issues

• Analysis of In-River Flood Measures on Flood Risk and Stream Ecosystem by Paired-Watershed Approach: a Case Study  
  by Hurem Dutal 
  Abstract: This study investigated the effects of small-sized in-river flood control structures on both flood susceptibility and stream ecosystems using the paired-watershed approach in Turkiye. Differences in daily streamflows among the pre-treatment, treatment, and post-treatment periods were determined using analysis of variance (ANOVA). The effects of flood control structures on the stream ecosystem were evaluated using the low flow and flashiness indicators, whereas their impact on flood susceptibility was assessed using high flow. The results showed that flood susceptibility increased by 19.8%, while low flows increased by 13.4% and flashiness decreased by 6% after the flood control measures.
  Keywords: flood risk; small-sized in-river structures; weir; paired-watershed; streamflow; low flows; Türkiye.
  DOI: 10.1504/IJGW.2026.10076016
   
  
• Optimisation of Multi-Source Meteorological Data Fusion Algorithms Based on Artificial Intelligence  
  by Yingkui Yang, Hongliang Han, Chao Li, Yang Dong 
  Abstract: Frequent extreme weather events demand high-resolution forecasts, yet current multi-source fusion models struggle with small-scale accuracy. This study aligns multi-source meteorological data via timestamp synchronization and spatial interpolation, then employs a multi-scale CNN to capture both local precipitation disturbances (small kernels) and large-scale atmospheric circulation (large kernels). A self-attention mechanism strengthens extreme weather signals, while a GNN-based fusion subnetwork models cross-source associations. A multi-task loss further enhances prediction accuracy and spatiotemporal consistency. Experiments show clear advantages: the proposed method achieves lower MAE (0.25 vs. 0.68, 0.49, 0.35 for baselines), higher spatial consistency (0.96), extreme event recognition (>0.9), and strong correlation (0.983), demonstrating its effectiveness in high-resolution small-scale extreme weather prediction.
  Keywords: Multi-Source Meteorological Data Fusion; Artificial Intelligence; Deep Convolutional Network; Attention Mechanism; Extreme Weather Prediction.
  DOI: 10.1504/IJGW.2026.10076108
   
  
• A Study on Optimising the Green Coverage of Low-Carbon Building Landscapes Using Deep Learning Algorithms  
  by Yuyao Zhang, Jiahao Zhang 
  Abstract: This study proposes a boundary-cross supervised segmentation network (BCS-SegNet) for accurate green coverage segmentation in urban street view images. Integrating decoupled residual self-attention and boundary cross-supervision, BCS-SegNet outperforms models like R-CNN, U-Net, and TransUNet in mIoU and dice coefficients. The method supports low-carbon building landscape optimisation by precisely quantifying green view rates, offering a robust tool for urban green planning.
  Keywords: Deep Learning; Building landscapes; SegNet; Green visual index.
  DOI: 10.1504/IJGW.2026.10076109
   
  
• Low Carbon Planning Model for the Development of Distributed Green Energy Industry Under the Background of Urban Digital Transformation  
  by Wei Liu, Wenxia Tong 
  Abstract: This paper proposes a digital twin-driven carbon flow dynamic planning framework (DT-CFDP) to align multi-source real-time data with static low-carbon urban planning. Integrating city information modelling (CIM), real-time data fusion, proximal policy optimisation and extended Kalman filter-based calibration, DT-CFDP enables 5-minute adaptive energy scheduling under carbon intensity constraints. Validated on five pilot cities (20202023), it reduces carbon emissions per unit GDP by 32.34% versus LEAP, MPC, and DQN baselines, improves response time by 32.8%, reduces lifecycle costs by 18.7%, and maintains 69.30% median renewable penetration, offering a real-time, adaptive low-carbon planning approach for urban green energy systems.
  Keywords: Smart City; Renewable Energy; Low Carbon Planning; Digital Twin; Reinforcement Learning.
  DOI: 10.1504/IJGW.2026.10076618
   
  
• Evaluation of Porous Aluminum Materials Used in Crash Boxes in Automobiles with AHP-BORDA in terms of Sustainable Production  
  by Samet K?rm?z?tepe, Nil Toplan, Alparslan Serhat Demir 
  Abstract: Today, the increase in global warming caused the automotive industry to develop more sustainable, lightweight and safe components, leading to rise in environmental concerns. Crash boxes, especially in the front the vehicle parts, are critical to ensuring passenger safety, absorbing energy in the event of an impact. In this study, aluminum-based porous materials were evaluated in terms of sustainable production for use in crash boxes using the Analytical Hierarchy Process (AHP) and BORDA methods. The results obtained by the AHP-BORDA methods reveal that Sintered Porous Aluminum was determined as the most suitable material in terms of environmental sustainability.
  Keywords: Sustainable Production; Emission;Automotive;Crash Boxes;AHP;BORDA.
  DOI: 10.1504/IJGW.2026.10076829
   
  
• Accurate Calculation Model for Carbon Sequestration Efficiency of Urban Green Spaces by Integrating Lidar and Multispectral Images  
  by Qianhe Xiang 
  Abstract: Accurately measuring urban green space carbon sequestration efficiency is challenging due to the limitations of single remote sensing sources. This paper proposes the dual-modality-collaborative sensing network (DM-CSN), integrating LiDAR and multispectral imagery. Using CNNs for feature extraction and a cross-modal transformer attention module for dynamic fusion, the model optimises contribution weights for carbon sink estimation. Experimental results demonstrate high precision, with an R2 of 0.89 and RMSE of 0.17 kg C/(m2
  Keywords: LiDAR-Multispectral Fusion; Urban Green Space; Carbon Sequestration Efficiency; DM-CSN Model; Cross-Modal Attention.
  DOI: 10.1504/IJGW.2026.10076839
   
  
• Towards Carbon Neutrality in Higher Education: Carbon Emission Accounting and Mitigation Pathways in a Chinese University Campus  
  by Zifan Cheng, Yufei Yang Chen, Yu Xiao, Tingting Hu, Jingqi Deng, Yusen Duan, Min Zhao, Dungang Gu, Jiaqi Lu 
  Abstract: This study develops a campus-scale carbon accounting framework based on the emission factor method and applies it to a representative university campus in Shanghai, China, to quantify emission sources and assess decarbonisation pathways. Total emissions in 2023 amounted to 14625.65 tCO2-eq, primarily from electricity, natural gas, and wastes. Existing mitigation measures delivered 4035.21 tCO2-eq of carbon offsets through vegetation sinks, recycling, and rooftop photovoltaics. Scenario analyses integrating campus-level mitigation measures with power-system decarbonisation pathways reveal strong synergies among coordinated interventions, enabling a feasible transition toward low- and zero-carbon campuses. The framework provides transferable insights for university carbon neutrality strategies and contributes to broader decarbonisation efforts in the urban environment.
  Keywords: Campus carbon emissions; Carbon emission accounting; Carbon neutrality pathways; Carbon offset; Decarbonisation of power grid; Emission factor method; University sustainability.
  DOI: 10.1504/IJGW.2026.10077156
   
  
• Fuel Consumption and CO2 Emission Comparison of Conventional and Hybrid Vehicles Using Data-Based Models and ECMS  
  by Zeynep Burcu Acunas, Ovun Isin, Ozgun Balci 
  Abstract: In this study, data was obtained via the OBD2 output of a sample hybrid vehicle. Then, electric motor, battery, and internal combustion engine (ICE) maps were created. Vehicle fuel consumption and CO2 emission results over NEDC cycle and a local route were compared between conventional vehicle model equipped with spark ignition engine and hybrid vehicle model with ECMS energy management, developed in Matlab/Simulink environment. Vehicle fuel consumption decreased by 7.59% with the ECMS controlled hybrid system over the NEDC cycle, and by 12.03% under the local route conditions. Similar significant improvements obtained in CO2 emission results as well.
  Keywords: Hybrid vehicle; ECMS; CO2; fuel consumption.
  DOI: 10.1504/IJGW.2026.10077159
   
  
• Heterogeneous Dynamic Relationship between Urban Expansion and Carbon Emissions in China  
  by Xiaoqiang Zheng, Mengyue Li 
  Abstract: Urban expansion increases carbon emissions, but its dynamic relationship under heterogeneous conditions remains unclear. This study discusses the drivers and decoupling paths of carbon emissions in cities with different expansion types. The analysis indicates that the contribution levels and offset effects of the driving factors explained the carbon emission differences among city types, with faster urban expansion amplifying these effects. The level of economic development was the primary factor that inhibited the decoupling across different city types, while output carbon intensity and per capita GDP were the key drivers. These insights support tailored low-carbon strategies and collaborative mitigation efforts.
  Keywords: urban expansion; carbon emissions; Theil index; decomposition analysis; decoupling analysis.
  DOI: 10.1504/IJGW.2026.10077196
   
  
• Assessment of Environmental Impacts in Travertine Mining with Life Cycle Approach  
  by Beyzanur Uzuntaş, Burcu Onat, Murat Yilmaz 
  Abstract: Natural stones like travertine play a vital role in construction, cladding, and architecture, contributing to both economic development and sustainability. This study evaluates the environmental impacts of producing 1 m3 of travertine blocks using life cycle assessment (LCA) at a quarry in Denizli, western Turkiye. The system boundary was defined from cradle to gate, including extraction, cutting, and transportation stages. The LCA was conducted using GaBi Education 8.0 and the CML-2001 method, covering eleven impact categories, including global warming potential (GWP), acidification potential (AP), and eutrophication potential (EP). The GWP for producing 1 m3 of travertine was 32.6 kg CO2 eq, primarily due to electricity use during cutting and diesel consumption in transport. Scenarios using solar and wind energy reduced the GWP to 7.18 and 6.03 kg CO2 eq, respectively, demonstrating the strong potential of renewable energy to minimise environmental impacts in the natural stone industry.
  Keywords: Natural stone mining; travertine; life cycle assessment; environmental impact.
  DOI: 10.1504/IJGW.2026.10077321
   
  
• Farmland and plane (Platanus orientalis L.) stand accumulated more phytolith-occluded organic carbon in the ancient Yellow River channel  
  by Xinru Gu, Baoxian Tao, Yuqing Jiang, Shijun Zhu 
  Abstract: Response of phytolith-occluded organic carbon (PhytOC) to different land use types remains unclear. Since the 1950s, the ancient Yellow River channel (ACYR) has been converted into farmland (FL), forest (dominated by species such as Sophora japonica L, Populus nigra, Fraxinus chinensis, and Platanus orientalis L. (PO), etc), and orchard. This study firstly identified the land use types associated with the greatest soil PhytOC storage in the AYRC. Results showed that FL and PO exhibited the largest soil PhytOC storage. Firstly, the substantial PhytOC storage in the aboveground litter layer of FL and PO enhanced the return of PhytOC to the soil. Secondly, elevated concentrations of soil Si fractions in FL and PO may promote the phytolith production, while increased concentrations of soil amorphous iron oxide may inhibit the phytolith dissolution. Overall, these findings suggest that FL and PO are conducive to the PhytOC accumulation in the ACYR.
  Keywords: Phytolith-occluded organic carbon; soil Si fractions; land use types; sandy land; ancient Yellow River channel.
  DOI: 10.1504/IJGW.2026.10077414
   
  
• Enhancing Rainfall Data Homogenisation in Mountainous Basins, Vietnam: a Case Study of Da River Basin  
  by Tran Khac Thac, Nguyen Tien Thanh, Nguyen Hoang Son, Vu Thi Minh Hue 
  Abstract: Reliable rainfall data are essential for hydrological studies in mountainous basins with sparse monitoring networks such as the Da River Basin (DRB). This study evaluates the homogeneity and consistency of long-term rainfall records from 16 stations over the period 19812022. Double mass curves and six statistical homogeneity tests were applied at a 95% confidence level, followed by data adjustment using quantile mapping. Results show that 25% of stations exhibit significant inhomogeneity. The adjustment effectively improves data consistency, providing a reliable rainfall dataset for hydrological and climate-related applications in the DRB.
  Keywords: Homogeneity testing; DRB; quantile mapping.
  DOI: 10.1504/IJGW.2026.10077426
   
  
• Embodied Carbon Emissions Patterns and Industrial Chain Pathways in the Beijing-Tianjin-Hebei Region  
  by Juan Tan, Jinyu Wei 
  Abstract: This paper analyzes embodied carbon flows across the Beijing-Tianjin-Hebei region using an environmentally extended multi-regional input-output model coupled with structural path analysis. By tracking CO2 emissions across seven aggregate sectors during 20172020, this study reveals a pronounced production-consumption asymmetry: production-based emissions increased by 153.2 Mt CO2 while consumption-based emissions decreased by 75.5 Mt, with Hebei identified as the principal net carbon exporter. Traditional heavy industry accounts for over 50% of embodied transfers, and approximately two-thirds of emissions originate in first-stage supply chains. An inter-regional shared accountability framework for coordinated emission reduction is proposed.
  Keywords: embodied carbon; MRIO; structural path analysis; supply-chain emissions; industrial transformation; regional climate governance.
  DOI: 10.1504/IJGW.2026.10077445
   
  
• Time series forecasting of global temperature anomalies using machine learning  
  by Ravi Patel, Aditya Kumar, Jainath Yadav, Mrityunjay Singh 
  Abstract: Accurate forecasting of global temperature anomalies is essential for understanding long-term climate variability and guiding climate change policies. This study compares classical machine learning (ML) and deep learning (DL) models for predicting global mean temperature anomalies using data from 1880-2024. The framework incorporates extensive pre-processing with several ML and DL architectures. Results show that ML models, especially SVR and Lasso, achieve higher accuracy and efficiency than DL approaches. Forecasts for 2025-2034 indicate a consistent warming trend, with DL models projecting anomalies surpassing 1.5 C by 2031, while regression-based methods estimate more moderate increases of 1.1-1.3C .
  Keywords: global temperature anomalies; climate forecasting; machine learning; deep learning; time series prediction; climate change.
  DOI: 10.1504/IJGW.2026.10078018
   
  
• Indoor environment design of university classrooms based on improved genetic algorithm and lighting conditions  
  by Wenting Chang, Zhaofeng Wang, He Dong 
  Abstract: University classrooms face challenges such as insufficient coordination between natural and artificial lighting, glare, and uneven illuminance. This study deploys sensors to collect light and heat data, and conducts parametric modelling based on Grasshopper. Ladybug simulates daylight performance indicators and uses an improved genetic algorithm for multi-objective optimisation to improve overall daylight environment performance. The results showed that after optimisation, in classrooms with 5070 people, the DA increased from 62.3% to 72.5%, and the DGP decreased from 0.52 to 0.44. In classrooms with 120150 people, the DA improved by 9.75%. These findings demonstrate that the improved genetic algorithm enhances daylight environment performance, providing a basis and pathway for design.
  Keywords: improved genetic algorithm; indoor lighting environment design; university classrooms; multi-objective optimisation; dynamic lighting index.
  DOI: 10.1504/IJGW.2026.10078113
   
  
• Impacts of temperature-driven climate change on agricultural pest population dynamics: neural network-based simulation and risk prediction  
  by Hongfei Zhang, Zhengbing Wang, Mingsheng Yang, Yan Zhang, Zhen Li 
  Abstract: Global climate change alters agricultural pest population dynamics. Existing neural network prediction models lack biological interpretability. This study constructs the BIT-Net framework integrating biological mechanisms and data-driven channels to simulate the temperature-pest dynamic relationship. Experiments in the North China Plain demonstrate a high prediction accuracy. The framework identifies the optimum developmental temperature and the high-temperature inhibition threshold. The risk warning system provides advance warnings for multiple climate scenarios. Cross-regional migration tests confirm the framework maintains high prediction accuracy across varying conditions.
  Keywords: climate change; agricultural pests; population dynamics; neural network; risk prediction.
  DOI: 10.1504/IJGW.2026.10078114
   
  
• Machine learning modelling of coordinated development of environment and economy in climate change  
  by Wenjuan Li, Ming-Hung Shu, Jui-Chan Huang, Chih-Chin Kuo 
  Abstract: Traditional linear models fail to capture policy synergies between environment and economy, leading to low predictive accuracy. This paper proposes a machine learning framework for coordinated development of the environment and the economy (CDEE). It employs a bidirectional long short-term memory (Bi-LSTM) to capture long-term dependencies in key indicators, such as carbon emission intensity. The extreme gradient boosting-light gradient boosting machine (XGBoost-LightGBM) model combines XGBoosts feature filtering with LightGBMs high-dimensional data processing. Bi-LSTM achieved a root mean square error (RMSE) of 0.82, while XGBoost-LightGBM had a variance of only 0.0001 for environment-economy coordination index (EECI).
  Keywords: climate change; environment and economy; collaborative development; machine learning; extreme gradient boosting; XGBoost; bidirectional long short-term memory; BiLSTM.
  DOI: 10.1504/IJGW.2026.10078115
   
  
• Intelligent assessment of low-carbon landscape design effects based on transformer and temporal point cloud analysis  
  by Shu Wang 
  Abstract: Amid urbanisation and climate change, low-carbon landscape design is vital for sustainability, yet traditional assessments rely on subjective, static methods. This study introduces a transformer-based framework analysing temporal point clouds to intelligently quantify carbon sequestration, urban heat island mitigation, and biodiversity. Validated on multi-temporal LiDAR datasets from urban parks, our model achieves 92.5% carbon accuracy (18.7% improvement over baselines), thermal comfort prediction RMSE of 0.42, and biodiversity correlation of 0.89. The results demonstrate significant advancements in accuracy and interpretability for dynamic landscape performance evaluation, advancing AI applications in sustainable urban design.
  Keywords: low-carbon landscape design; temporal point cloud; transformer; spatiotemporal feature extraction; carbon sequestration assessment; urban heat island mitigation; biodiversity evaluation; intelligent assessment.
  DOI: 10.1504/IJGW.2026.10078184
   
  
• Advanced machine learning framework for predicting municipal solid waste generation trends toward resilient and environmentally sustainable urban development  
  by Ankit Bansal, Umang Soni, Aqueel Ahmad 
  Abstract: Due to the rapid urbanisation and industrial growth, municipal solid waste (MSW) generation becomes a global challenge. The present study applies machine learning (ML)-based prediction models to estimate MSW generation trends in Singapore spanning the period 2003-2020. Model performance was evaluated using the various error metrics, while variable importance was assessed via the Pearson correlation coefficient (PCC) and partial dependence plots. The findings demonstrates that ensemble-modelling approaches offers high predictive accuracy and can be generalised to various regions for effective MSW forecasting, contributing to environmental resilience in line with sustainable development goals (SDG) 12 and 13.
  Keywords: municipal solid waste; MSW; waste generation prediction; machine learning; sustainable urban development; environmental sustainability.
  DOI: 10.1504/IJGW.2026.10078191
   
  
• Climate Mitigation Implications of Small Modular Reactor-Integrated Microgrids in Baseload Power-Centric Smart Cities  
  by T.A.E. H.O. Woo, Chang Hyun Baek, Kyung B.A.E. Jang 
  Abstract: This study uses the System Dynamics method with Vensim to analyze microgrids integrating fossil fuels, Small modular reactor-based nuclear, and renewable energy for climate mitigation. Across four 50-year scenarios, Current2 shows the strongest results - its Power Grid rises nearly 800 times, and Smart City Integrity increases over 200 times, making it most suitable for smart city applications. Statistical analysis confirms that Current2 has the highest variability. As AI-driven industries increase power demand, nuclear energy becomes increasingly vital to stabilize supply and complement intermittent renewables, supporting sustainable energy transition and climate change response.
  Keywords: Climate Mitigation; Microgrid; Baseload; Energy; Small Modular Reactor (SMR).
  DOI: 10.1504/IJGW.2026.10078292
   
  
• Impact of Climate Change on the Potential for Extreme Rainfall Events in the Lesser Sunda Islands - Indonesia  
  by Aura Syakira Fidinina, Alvin Pratama, Amalia Nurlatifah 
  Abstract: Climate change strongly influences the intensity and distribution of extreme rainfall in tropical regions, including the Lesser Sunda Islands This study examines historical trends (19842014) and future projections (20152060) using CHIRPS data and the ACCESS-ESM1.5 model (NEX-GDDP-CMIP6), corrected with quantile mapping. Post-correction evaluation indicates improved performance, with correlation above 0.7 and reduced RMSE. Extreme rainfall was assessed using ETCCDI indices (RX1day, R50mm, and CDD) and SPI-3 for seasonal drought. Projections show increases RX1day and R50mm under SSP1-2.6 and SSP2-4.5, while CDD increases across all scenarios, reaching ~180 days under SSP3-7.0. SPI1.0 also increases, indicating more frequent and widespread droughts in the future.
  Keywords: Extreme rainfall; ETCCDI; Lesser Sunda Island; NEX-GDDP-CMIP6; Quantile Mapping; SPI-3.
  DOI: 10.1504/IJGW.2026.10078318
   
  
• Dynamic Evolution Analysis of River Ecological Economic Belt at Regional and Urban Scales  
  by Yeming Lyu, Yuxiao Shang, Zhenghui Chen, Yihui Xu 
  Abstract: Taking the Huaihe River Basin as a case study, this study examines the dynamic changes in water resources resulting from global urbanization and economic development from 2010 to 2024. This study develops an adaptive evaluation index system that includes social, economic, and ecological dimensions. The results showed that the adaptability of social systems has been enhanced, the adaptability of ecosystems has remained stable but with significant spatial differences, and the overall adaptability has improved. Research has shown that the dynamic balance between society and ecosystems is crucial for the adaptive stability of watersheds.
  Keywords: Huaihe River basin; Ecological economic belt; Dynamic evolution; Adaptive evaluation; Social system.
  DOI: 10.1504/IJGW.2026.10078465
   
  
• An Integrated Modelling Approach for Sustainable Development of the UNESCO classified ecosystem (Ichkeul Lake, North Africa)
  by Liping Ouyang, Zhirui Huang, Enen Chen, Ying Bai 
  Abstract: This paper presents the results of a simulation study on the impact of integrating water intake and output for and from Lake Ichkeul, a Ramsar and UNESCO reserve wetland in North Tunisia. Three scenarios were simulated over nine years, with varying amounts of freshwater injected into the lake from different dams. The study showed that a non-active lake management policy would result in severe ecosystem degradation, with the lake eventually becoming a salt marsh. Under the status quo, the area and density of Potamogeton would decrease dramatically and the number of migratory birds would rapidly decline. The second scenario would allow the ecosystem to be barely resilient and maintain itself, while the third scenario would lead to long-term sustainability and stabilize all ecosystem components. The results of the present study have implications for water management policies and the preservation of the unique biodiversity of the Ichkeul Lake ecosystem.
  Keywords: Ichkeul Lake; Water management; Ecosystem resilience; Wetland conservation; Scenarios.
  
  
• Collaborative System of Low-Carbon Sustainable Energy and Grassroots Governance Based on Big Data Analysis  
  by Binghui Lei, Peng Xu 
  Abstract: Low-carbon energy operation at the grassroots level suffers from disconnected energy-governance data, feedback lag, and no computable coupling mechanism. This study constructs a collaborative system based on big data, forming a heterogeneous temporal graph through unified mapping and time alignment, and introducing attention-driven dynamic graph representation learning to model energy and governance behaviour evolution. Low-carbon constraints and policy gradient updates drive energy scheduling and governance incentives within a shared closed loop. Experiments show carbon emission intensity stabilises at 0.39 kgCO2kWh, supply-demand deviation at 2.39%, and governance response delay at 1.8 time windows, with a collaborative stability index converging to 0.87.
  Keywords: Big Data Analysis; Low-Carbon Sustainable Energy; Grassroots Governance; Dynamic Graph Representation Learning; Collaborative Optimization.
  DOI: 10.1504/IJGW.2026.10078595
   
  
• Energy-daylight interaction perspective: effects of natural shading, photosensor and orientation  
  by Yasin Kaya, Kadir Zengin, Sebahattin Ünalan 
  Abstract: Buildings account for a substantial share of global energy consumption and related carbon emissions, making the building sector a key component of global warming mitigation strategies. Therefore, while improving energy efficiency remains highly important, this study additionally investigates daylight comfort criteria, which also play a critical role in sustainable building design.In this study, the combined effects of natural shading, daylight-responsive photosensors, and building orientation on energy consumption and daylighting performance were comprehensively analyzed using a reference building model. These parameters were examined under different scenarios through comparative energy assessments that explicitly considered heat transfer processes.The main findings indicate that integrating photosensors with optimized building orientation significantly improves daylight comfort. Consequently, total energy consumption was reduced by up to 6.3% (4.11 MWh/year), accompanied by a 7.6% reduction in CO? emissions (2,175 kgCO?/year).
  Keywords: building; energy-daylight interaction; natural shading; photosensor; orientation.
  DOI: 10.1504/IJGW.2026.10078658
   
  
• Decarbonizing Algeria's Phosphate Industry: A Life Cycle Assessment of Environmental Hotspots and Policy-Ready Solutions  
  by Ali Makhlouf, Hamza Cheniti, Hani Amir Aouissi, Marc Azab 
  Abstract: Algeria's phosphate sector faces the challenge of increasing production while limiting environmental impacts in a water-scarce and fossil-fuel-dependent context. We applied a cradle-to-gate Life Cycle Assessment using the ReCiPe2016 method to evaluate the environmental footprint of beneficiated phosphate. Calcination accounts for 58% of greenhouse gas emissions, while diesel-based logistics represent 47.09% of fuel consumption. Scenario analysis indicates that electrified conveyors could reduce emissions by 6.83% and crusher relocation can lower transport impacts. Although Algeria shows lower energy intensity than Morocco and Tunisia, its water recycling rate remains limited. Sustainable pathways include renewable energy integration, improved water circularity, and SDGs alignment.
  Keywords: Algeria; Electrified conveyor systems; Groundwater depletion; LCA; Phosphate production; Renewable energy integration; SDGs.
  DOI: 10.1504/IJGW.2026.10078662
   
  
• Machine Learning Simulation Analysis of the Impact of Extreme Environmental Events Driven by Global Warming on Economic Dynamics  
  by Zhimin Wu, Ye Tian 
  Abstract: This paper proposes an XGBoost-SHAP framework to model nonlinear climate-economy linkages under extreme events. Using multi-source panel data, it constructs a hierarchical intensity index for complex extremes. Bayesian-optimised XGBoost minimises cross-validated MSE to capture climate shocks nonlinear economic impacts. TreeSHAP yields global feature rankings and local attributions, revealing key drivers: high temperature (SHAP = 0.42), precipitation frequency (0.38), and long-term drought (0.35). In agriculture-heavy economies (> 40% share), extreme heat suppresses annual growth by 0.57 on average. The model achieves RMSE 1.52.1 and MAPE 4.8%8.2%, offering high accuracy and interpretability for differentiated climate adaptation policy.
  Keywords: Climate Warming; Extreme Environmental Events; Extreme Gradient Boosting; SHapley Additive exPlanations; Nonlinear Modeling.
  DOI: 10.1504/IJGW.2026.10078664
   
  
• Spatiotemporal Evolution and Multi-Scenario Simulation of Urban Green Space Carbon Sinks in the Guangdong-Hong Kong-Macao Greater Bay Area from Accounting to Prediction  
  by Liping Ouyang, Zhirui Huang, Enen Chen, Ying Bai 
  Abstract: This study couples the InVEST and PLUS models to analyse spatiotemporal patterns and future carbon stock in the Guangdong-Hong Kong-Macao Greater Bay Area from 1995 to 2050 under natural, ecological and economic scenarios. Results show that carbon stock decreased from 1995 to 2025, with a high periphery and low centre pattern, primarily due to the conversion of arable and forest land to construction. Ecological protection increases future stocks while economic development accelerates loss. The study provides a quantitative basis for land optimisation and carbon neutrality pathways.
  Keywords: Urban Green Space Carbon Sinks; Spatiotemporal Evolution; Multi-Scenario Simulation; Guangdong-Hong Kong-Macao Greater Bay Area; InVEST-PLUS Model Coupling.
  DOI: 10.1504/IJGW.2026.10078803
   
  
• Legal Liability under Mode of Rural Environmental Pollution Control Based on Internet Technology  
  by Ting Yuan 
  Abstract: As a major agricultural nation, China faces rural environmental pollution control as a critical issue for agricultural and rural development. In the context of the rural revitalization strategy, rural environmental protection has become especially urgent. This study takes rural environmental improvement as its starting point, leverages Internet technology to develop a new online trial system for rural environmental pollution cases, and refines the legal liability framework for pollution control. Using core principles of game theory, the study constructs a dynamic tripartite game model involving enterprises, environmental regulators, and villagers, derives the equilibrium solution, and explores corresponding legal countermeasures. Experimental results demonstrate that implementing the Internet-based case trial system in rural environmental pollution governance significantly increases villagers' satisfaction by 9.63% and effectively advances rural pollution control efforts.
  Keywords: Rural Environmental Pollution Control; Legal Liability Issues; Internet Technology; Game Theory.
  DOI: 10.1504/IJGW.2026.10078805
   
  
• Tracking the Effectiveness of Marine Ecological Assessment and Restoration Around Nuclear Power Plants Based on Deep Learning  
  by Zhixian Wang, Depeng Li 
  Abstract: This deep learning system assesses marine ecosystem restoration around nuclear plants. Multi-source data form spatiotemporal tensors. A spatio-temporal convolutional neural network-transformer extracts parameters like chlorophyll-a. A graph neural network models ecological connections. Attention long short-term memory integrates data to quantify contributions. Experiments show: root mean square error of 0.85 and Nash-Sutcliffe model efficiency coefficient of 0.92 for chlorophyll-a; intersection over union of 0.405 for sea surface temperature anomalies; suspended matter prediction errors of 3.36.74%. The model enables high-precision dynamic assessment and reliable restoration tracking.
  Keywords: Deep Learning; Marine Ecological Assessment; Nuclear Power Plant Environment; Ecological Restoration Effectiveness Tracking; Spatio-Temporal Convolutional Neural Network.
  DOI: 10.1504/IJGW.2026.10078806
   
  
• Analysis of Digital Marketing Strategies in Enterprise Supply Chain under the Background of Low Carbon Economy  
  by Xiaoyu Yan 
  Abstract: To study the digital marketing strategies of enterprises under the background of low-carbon economy, this paper proposes a fuzzy neural network to evaluate the digital marketing of enterprises. Evaluation is the foundation for ensuring the effective operation of enterprise work methods. Scientific, effective, and fair evaluation can stimulate the enthusiasm of enterprises. On the other hand, it can also improve the digital marketing performance of enterprises. The experimental results showed that 68 people believed shopping was very convenient under traditional marketing, while 160 people believed shopping was very convenient under digital marketing. Only 55 people believed that the service quality of traditional marketing was high, but 189 people believed that the service quality of digital marketing was high. It can be seen that under digital marketing, not only is shopping convenient, but people are also very satisfied with the service quality of digital marketing.
  Keywords: Low-carbon Economy; Fuzzy Neural Network; Enterprise Digital Marketing; Enterprise Supply Chain; Enterprise Evaluation.
  DOI: 10.1504/IJGW.2026.10078809
   
  
• Assessment of Decadal Rainfall Prediction Skill Across Southeast Asia Using CMIP6 Models  
  by Dara Kasihairani, Supari Supari, Rahmat Hidayat 
  Abstract: This study evaluates decadal climate prediction (DCP) skill over Southeast Asia using CMIP6 hindcasts initialised during 19602010. Rainfall prediction skill is assessed for winter (DJF) and summer (JJA) using the anomaly correlation coefficient (ACC), root mean square error (RMSE), and normalised standard deviation (NSD). Results show spatial clustering into three subregions (MCSEA, MLSEA, and PLSEA). Domain-averaged ACC remains weak (0.01 to 0.07) but exhibits substantial spatial variability (0.5 to 0.5). RMSE is typically 140160 mm. Skill is higher at early lead times (LY2) and in DJF, decreases at intermediate leads, and partially recovers at longer leads (LY8LY9).
  Keywords: forecast skill; decadal prediction; region; lead time; ensemble model.
  DOI: 10.1504/IJGW.2026.10078906