38796 research outputs found
Sort by
Mapping the transition of the EU steel industry to carbon neutrality
Get PDFThis factsheet provides an overview of sectoral emission sources, emissions breakdowns, decarbonisation trajectories, and estimated technology-specific CO₂ abatement costs. It further examines the evolution of decarbonisation technology maturity (from research and innovation to demonstration and deployment) in the timeline from 2025 to 2050 and evaluates the extent to which this evolution aligns with relevant policy targets and objectives.JRC.C.7 - Energy Transition Insights for Polic
Novel deep learning algorithm in soil erodibility factor predicting at a continental scale
No full textSoil erosion poses significant environmental and economic challenges, adversely affecting soil fertility and global agricultural productivity. We developed a novel model based on the Multi-Head Squeeze-and-Excitation Residual One-Dimensional Convolutional Neural Network (MH-SE-Res1DNet) to predict the soil erodibility factor (K) across Europe, representing the first application of this model for such a purpose worldwide. We conducted a comparative analysis using five benchmark machine learning algorithms, i.e., Random Forest (RF), Artificial Neural Network–Multilayer Perceptron (ANN-MLP), Support Vector Regression (SVR), Alternating Model Tree (AMT), and Pace Regression (PR), to assess the efficacy of our model. The results showed that the MH-SE-Res1DNet deep learning model had an outstanding ability for the K prediction. The model's lowest error (MAE = 0.0025, RMSE = 0.0031) and highest coefficient of determination (R2 = 0.943) were attained during the validation phase. Benchmark models demonstrated lower performance compared to the MH-SE-Res1DNet model, with R2 values ranging from 0.880 to 0.912 and slightly higher errors across MAE and RMSE metrics. The sensitivity analysis of MH-SE-Res1DNet showed that its performance depends predominantly on key soil factors, particularly topsoil texture (M) and organic matter (OM) concentration. This model establishes a data-driven framework that significantly advances soil erodibility prediction by leveraging machine learning. It surpasses traditional methods and existing machine learning approaches in accuracy, efficiency, and scalability, setting a new benchmark for soil conservation planning and enabling adaptable, evidence-based land management strategies across Europe and worldwide.JRC.D.1 - Land and Climat
Deep Learning-Based Reduced-Order Methods for Fast Transient Dynamics
No full textIn recent years, large-scale numerical simulations played an essential role in estimating the effects of explosion events in urban environments, for the purpose of ensuring the security and safety of cities. Such simulations are computationally expensive and, often, the time taken for one single computation is large and does not permit parametric studies. The aim of this work is therefore to facilitate real-time and multi-query calculations by employing a non-intrusive Reduced Order Method (ROM).
We propose a deep learning-based (DL) ROM scheme able to deal with fast transient dynamics. In the case of blast waves, the parametrised PDEs are time-dependent and non-linear. For such problems, the Proper Orthogonal Decomposition (POD), which relies on a linear superposition of modes, cannot approximate the solutions efficiently.
The piecewise POD-DL scheme developed here is a local ROM based on time-domain partitioning and a first dimensionality reduction obtained through the POD. Autoencoders are used as a second and non-linear dimensionality reduction. The latent space obtained is then reconstructed from the time and parameter space through deep forward neural networks. The proposed scheme is applied to an example consisting of a blast wave propagating in air and impacting on the outside of a building. The efficiency of the deep learning-based ROM in approximating the time-dependent pressure field is shown.JRC.E.3 - Built Environmen
Bio-based plastics in a sustainable and circular bioeconomy
Get PDF‣ In the context of growing global demand of plastic products, bio-based plastics are generally included in the strategies to mitigate associated climate change and environmental impacts.
‣ Bio-based plastics include a broad category of polymers that can have a fossil-based counterpart or not, be biodegradable or not. In this brief, they are classified based on the step in the production process where fossil feedstocks are replaced by biomass.
‣ As of 2025, bio-based and biodegradable plastics accounted globally for roughly 0.5 % of the total plastics production. Global annual production capacity is around 2.3 Mt, and based on announced capacity expansions, it is projected to grow to about 4.7 Mt by 2030.
‣ The scale up of the sector is challenged by several issues, mainly related to competitiveness, sustainable feedstock sourcing and use, end of life management, assessment of environmental impacts, innovation aspects (e.g. technical development), and consumers’ attitudes.
‣ The bio-based plastic industry could potentially grow in the EU, relying on domestic feedstock. Such development is expected not only to bring socio-economic advantages in terms of job opportunities and strategic autonomy, but also to help improve recyclability and end-of-life management.
‣ Replacing fossil feedstocks generally lowers the greenhouse gas emissions over the product’s whole life cycle, while for the other environmental impact categories trade-offs may also occur.JRC.D.3 - Sustainable Supply Chains and Bioeconom
Review of the FIDELIO model: A New Generation of Dynamic, Econometric and General Equilibrium Input-Output Model
Get PDFThis technical report reviews the FIDELIO model, a dynamic, econometric, and general equilibrium input-output model developed by the Joint Research Centre (JRC) of the European Commission. The review was conducted by a panel of experts to assess the model's scientific rigour, data foundation, empirical implementation, and peer-review status, in line with the EU's Better Regulation Policy and the Inter-Institutional Agreement on Better Law Making.
The FIDELIO model is designed to estimate the socio-economic and environmental impacts of EU industrial, trade and innovation policies for a fair and sustainable EU economy, mostly oriented to demand policies but also flexible to incorporate supply policy shocks, too. It is built on a solid foundation of the FIGARO database, which is maintained by Eurostat, and aims to capture spillover and rebound effects to quantify impacts on jobs, growth, investments, resource use, emissions, and trade balance. The model has evolved through several versions, with the latest iteration featuring improvements in modularity, flexibility, and dynamic structure.
The review panel's general comments commend the model's comprehensive approach, which blurs the lines between traditional IO systems, econometric IO systems, and computable general equilibrium models. The panel appreciates the model's disaggregation of industries and households, which allows for granular analysis. However, the panel also points out the need for regular updates to the FIGARO database, transparent documentation of empirical implementation, and continuous exposure of the model to the scientific community and policymakers.
Specific comments from individual panel members address various aspects of the model, including the need for a clear statement of the EU policy objectives of the European Union where FIDELIO will have to be applied, the suitability of the model for contributing to well-being measures, and the importance of capturing the dynamics of competition and market power. The panel also suggests improvements in the treatment of technological change, flexibility, and substitutability, as well as the integration of financial accounts and regional integration, qualitative changes in the economy, demographic challenges, and micro-macro links within the model.
In response to the review, the FIDELIO team has proposed several lines of action to enhance the model's capabilities. These include updating the FIGARO database, refining parameter estimation and calibration, exploring new production functions, incorporating firm heterogeneity, and developing a FIDELIO community for collaborative development and use of the model. The team also plans to publish a flagship article on the model and maintain transparent documentation to facilitate knowledge transfer within the team.
The FIDELIO model is a critical tool for EU policy analysis, with the review providing valuable insights for its improvement. The proposed actions by the FIDELIO team aim to address the panel's recommendations and ensure the model. Overall, the FIDELIO model presents a comprehensive and transparent tool for analysing EU policy impacts in the European economy, with potential areas for further refinement and extensions to address evolving complexities and improve its credibility and applicability.JRC.B.7 - Innovation Policies and Economic Impac
Western Balkans participation in GVCs and FDI potential in the context of Smart Specialisation
Get PDFThis study examines the position of Western Balkan economies within global value chains (GVCs) and their capacity to attract foreign direct investment (FDI), linking the analysis with the broader framework of Smart Specialisation as a key pillar of the EU accession process and a strategic instrument for strengthening regional innovation ecosystems and deepening integration into the European Research Area. The results point to a steady increase in FDI between 2007 and 2022, accompanied by a growing role for services and knowledge-intensive activities, alongside differentiated trajectories of functional upgrading. Across multiple econometric specifications, the analysis shows that: (i) economic size, institutional proximity, EU integration and social connectedness are key drivers of international investment; (ii) obtaining EU candidate status increases FDI from EU countries by approximately 25–35% compared with non-EU sources; and (iii) while overall alignment with Smart Specialisation domains remains limited, it is more strongly driven by EU investors and reinforced by host-country R&D intensity and social connectedness. From a policy perspective, these findings suggest that more sophisticated economies - those more deeply embedded institutionally and relationally - are better positioned to attract strategic investment and support processes of functional upgrading and leverage Smart Specialisation as a mechanism for convergence within the EU integration process.JRC.B.3 - Territorial Developmen
The 2025 major fish kill on the Munster Blackwater, Ireland. Organisational responses and key recommendations.
Get PDFIn August 2025 a major fish kill event occurred on the Munster Blackwater, Ireland. It is estimated that up to 42,000 fish were killed. This was the largest recorded fish kill in Ireland and it caused widespread upset in the community and across the country. The initial response by state agencies was rapid with investigations occurring within hours of notification. However, the suspected cause and source of the pollution remains undetermined. The inter-agency report considered that the pollutant could have entered the river on the 5th or 6th of August and subsequently dissipated before notification of fish deaths on the 11th.
This report reviewed the response of state agencies and the inter-agency report produced. Key recommendations are aimed at preventing, detecting and improving the coordinated response to future fish kills. It is recommended to place continuous monitoring on major Irish rivers, prepare, agree and test a multi-agency plan for major fish kills to include comprehensive sampling of the river as well as discharging facilities in the catchment. A communication strategy should focus on immediately passing on knowledge, specifying any uncertainties, investigative steps being taken, and give immediate advice for the public. Research should determine high risk areas for fish kills for preventative action and to determine the economic cost of the fish kill. An intensification of ongoing restoration efforts is now needed in the catchment to achieve the environmental objectives under the Water Framework Directive.JRC.D.2 - Ocean and Wate
Residential load forecast: An enhanced machine learning model with socio-economic data and synthetic features
No full textAccurate forecasting of residential electricity loads is a critical step toward enabling demand-side flexibility and supporting efficient and reliable power system operation. This paper presents a novel Feature-Enhanced Residential Energy Forecast (FEREF) model for residential households that does not rely on historical load profiles, which are often unavailable. Instead, the proposed method bases its estimations on: (i) socio-economic data and building characteristics, (ii) synthetic features summarizing past energy behavior, and (iii) weather information. Following a detailed discussion on training data selection and synthetic feature design, the development and tuning of the proposed FEREF model, based on an Extra-Trees regressor, are described. The accuracy and sensitivity of the model are evaluated through representative case studies and benchmarked against alternative approaches from the literature. Results demonstrate that the FEREF model attains an R2 score of 0.82 and an average MAPE of 37.69%, demonstrating significant performance improvements with respect to the current state of the art, notably achieved while not requiring complete knowledge of historical load profiles. This makes the approach particularly suitable for newly connected households and data-scarce environments.JRC.C.3 - Energy Security, Distribution and Market
Gender differences in successor identification within family farms
No full textGrowing attention has been paid to gender balance in EU agriculture due to a persistently lower presence of female farm managers. While several studies have examined factors affecting successor selection in family farms, there is limited evidence on how the process of successor identification differs between male and female successors in the EU. This paper aims to address this gap while highlighting the factors that specifically support or hinder the choice of a female successor. Based on a survey of 559 farmers from Italy and Poland, we employ logistic and multinomial logistic regression to model the likelihood of identifying successors. Firstly, we find a significant difference in the factors affecting the identification of female and male successors, both in the pooled sample and in each country separately. Secondly, female successors are more likely to be identified than male successors in organic farms not belonging to producer organisations, where women are already involved in farm management and rely on the full incumbent's involvement in farming, stronger ties to public institutions, and certain retirement plans. In contrast, patrilineal family structures, such as having a firstborn son, multiple children or family members with health issues, tend to favour male successors. Although start-up support for young farmers may enhance the overall likelihood of successor identification, it does not appear to have a specific effect on the likelihood of selecting female rather than male successors. On this basis, the paper draws some considerations for gender-tailored policies to support gender balance in EU agriculture.JRC.D.4 - Economics of Food System
Federation of toxicological data resources for in silico new approach methodologies (NAMs)
No full textNext Generation Risk Assessment (NGRA) promotes animal-free, exposure-informed, and hypothesis-driven approaches to chemical safety assessment. In silico tools, such as quantitative structure-activity relationship (QSAR) models, are valuable new approach methodologies (NAMs) for use in NGRA. However, the practical implementation of in silico NAMs remains limited by challenges in data availability, heterogeneity, and regulatory acceptance. In this study, federated learning is introduced to advance chemical safety assessment while leveraging proprietary data domains. Federated learning is a decentralised machine learning approach where multiple organisations, devices or servers collaboratively train a model while keeping their data locally, sharing only model updates to preserve confidentiality and privacy. Three use cases were simulated with the Flower open-source federated learning framework, namely (i) federated analytics for dermal permeability (log Kp) screening; (ii) federated convolutional neural networks (CNNs) for mutagenicity prediction from SMILES strings, and (iii) federated eXtreme Gradient Boosting (XGBoost) models for predicting skin sensitisation potential using molecular fingerprints and descriptors. The results show that federated learning approaches can yield predictive performance comparable to centralised models while mitigating concerns over the visibility of, and access to, commercially sensitive data. Open challenges related to data curation, interpretability, and model governance, as well as future directions, are discussed. This work demonstrates that federated learning can facilitate secure collaboration across organisations, enhance the utility of distributed chemical datasets, and accelerate the adoption of in silico NAMs.JRC.F.3 - Systems Toxicolog