Fant 2768 publikasjoner. Viser side 6 av 277:
Evaluating the role of low-cost sensors in machine learning based European PM2.5 monitoring
We evaluate the added value of integrating validated Low-Cost Sensor (LCS) data into a Machine Learning (ML) framework for providing surface PM2.5 estimates over Central Europe at 1 km spatial resolution. The synergistic ML-based S-MESH (Satellite and ML-based Estimation of Surface air quality at High resolution) approach is extended, to incorporate LCS data through two strategies: using validated LCS data as a target variable (LCST) and as an input feature via an inverse distance weighted spatial convolution layer (LCSI). Both strategies are implemented within a stacked XGBoost model that ingests satellite-derived aerosol optical depth, meteorological variables, and CAMS (Copernicus Atmospheric Monitoring Service) regional forecasts. Model performance for 2021–2022 is evaluated against a baseline trained on air quality monitoring stations without any form of LCS integration. Our results indicate that the LCSI approach consistently outperforms both the baseline and LCST models, particularly in urban areas, with RMSE reductions of up to 15–20 %. It also exhibits higher accuracy than the CAMS regional interim reanalysis with a lower annual mean absolute error (MAE) of 2.68 μg/m3 compared to 3.32 μg/m3. SHapley Additive exPlanations based analysis indicates that LCSI information improves both spatial and temporal representativeness, with the LCSI strategy better capturing localized pollution dynamics. However, the LCSI's dependency on the spatial LCS layer limits its ability to capture inter-urban pollution transport in regions with sparse or no LCS data. These findings highlight the value of large-scale sensor networks in addressing spatial coverage gaps in official air quality monitoring stations and advancing high-resolution air quality modeling.
2026
Efficacy of individual and combined terrestrial and marine carbon dioxide removal
Abstract Limiting global temperature rise below 2°C requires significant reduction in greenhouse gas emissions and likely large-scale carbon dioxide removal (CDR). This study assesses the CO2 sequestration and efficacy of two CDR approaches, Bioenergy with Carbon Capture and Storage (BECCS) and Ocean Alkalinity Enhancement (OAE), applied individually and in combination. Using the Norwegian Earth System Model (NorESM2-LM), simulations were designed to ramp up deployment of BECCS and OAE, to an additional area of 5.2 million km² by 2100 for bioenergy feedstock for BECCS, and a CaO deployment rate of approximately 2.7 Gt/year for OAE within the exclusive economic zones of Europe, the United States and China. The combined land-ocean CDR simulation revealed a largely additive carbon removal effect. Over 2030-2100, OAE sequestered 7 ppm of CO 22 with an accumulated 82.3 Gt CaO, achieving a CDR effectiveness of 0.08 ppm (~ 0.17 PgC) per Gt CaO, while BECCS reduced 16 ppm of CO2, with CDR effectiveness of 3.1 ppm per million km² of bioenergy crops. Together, the carbon removal achieved by BECCS and OAE corresponds to anthropogenic CO₂ emissions of 5.4 Gt CO₂/year by 2100, slightly more than 60% of current global transport sector emissions. Notably, the efficiency of BECCS and OAE alone was unaffected by their concurrent deployment. Nevertheless, simulations revealed distinct non- linear interactions, such as declines in land and soil carbon sinks in the combined scenario. Furthermore, all simulations show negligible effects on the global annual mean temperature. These results highlight near-additive CDR responses even under net-negative emissions, but feedback on land and ocean carbon sinks must be considered when designing CDR portfolios. This study provides new insights into CDR portfolio design and Earth system feedback under an overshoot scenario, highlighting both their potential and the need for continued emissions cuts and supportive policies.
2026
New Approach Methodologies (NAMs) are gaining significant momentum globally to reduce animal testing and enhance the efficiency and human relevance of chemical safety assessment. Even with substantial EU commitment from regulatory agencies and the academic community, the full regulatory adoption of NAMs remains a distant prospect. This challenge is further complicated by the fact that the academic world, oriented toward NAMs development, and regulatory agencies, focused on practical application, frequently operate in separate spheres. Addressing this disconnect, the present paper, developed within the European Partnership for the Assessment of Risks from Chemicals (PARC), provides a clear overview of both the available non-animal tests and current evaluation practices for genotoxic and carcinogenic hazard assessment, while simultaneously highlighting existing regulatory needs, gaps, and challenges toward greater human health protection and the replacement of animal testing through NAMs adoption.
The analysis reveals a complex landscape: while the EU is deeply committed to developing and adopting NAMs, as outlined in its Chemical Strategy for Sustainability and supported by initiatives like PARC, prescriptive regulations such as Classification, Labelling and Packaging (CLP) and Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) still heavily mandate in vivo animal data for hazard classification, particularly for germ cell mutagenicity and carcinogenicity. This reliance creates a “too-short-blanket-problem,” where efforts to reduce animal testing may impact human health protection because of the current in vivo-based classification criteria. In contrast, sectors such as cosmetics and certain European Food Safety Authority (EFSA)-regulated products demonstrate greater flexibility toward progressive integration of NAMs. While the deep mechanistic understanding of genotoxicity and carcinogenicity has significantly advanced the integration of alternatives to animal tests into regulatory chemical hazard assessment, their broader and full implementation faces considerable challenges due to both scientific complexities (i.e., the development and validation of fit-for-purpose NAMs) and existing legislative provisions.
2026
Nitrogen dioxide (NO2) is a well-known air pollutant, mostly elevated by car traffic in cities. To date, small, reliable, cost-efficient multipollutant sensors with sufficient power and accuracy for community-based atmospheric studies are still lacking. The HAPADS (highly accurate and autonomous programmable platforms for providing air pollution data services) platforms, developed and tested in real conditions, can be a possible approach to solving this issue. The developed HAPADS platforms are equipped with three different NO2 sensors (7E4-NO2–5, SGX-7NO2, MICS-2711 MOS) and a combined ambient air temperature, humidity, and pressure sensor (BME280). The platforms were tested during the driving test, which was conducted across various roads, including highways, expressways, and national and regional routes, as well as major cities and the countryside, to analyse the environmental conditions as much as possible (Poland, 2024). The correlation coefficient r was more than 0.8, and RMSE (root mean squared error) was in the 3.3–4.3 μg/m3 range during the calibration process. The results obtained during the driving tests showed R2 of 0.9–1.0, which proves the ability of HAPADS platforms to work in the hard environmental conditions (including high rain and snow, as well as sun and a wide range of temperatures and humidity).
2026
Physics-Informed Deep Learning for Wind Downscaling over Oslo
Running a numerical weather model such as WRF at kilometre or sub-kilometre grid spacing over a regional domain is computationally expensive. We present physics-informed deeplearning models that ingest a single 9km WRF wind field and simultaneously predict two finer-scale wind fields at 3 km and 1 km resolution via dual decoder heads. Four representative architectures are benchmarked-Deep Residual U-Net (DeepRU), DEVINE, a bespoke 3-D Transformer, and a Fourier Neural Operator (FNO)-each trained with divergence-free, vorticity, and Navier-Stokes residual constraints plus Charbonnier and gradient perceptual losses. We train and validate our models on the city of Oslo for the year 2018. DeepRU achieves R2=0.94 (RMSE =0.050) at 3km and R2=0.89(RMSE=0.065) at 1 km. DEVINE, Transformer 3-D, and FNO yield 3 km scores of 0.91−0.93, with 1km scores lower by 0.02−0.08, illustrating the increased difficulty of finer-scale reconstruction. Physicsinformed losses improve all models compared to MSE-only baselines, and the residual architecture (DeepRU) remains most effective for this dual-scale task.
2025
Characterization of German SF6 Emissions
Sulfur hexafluoride (SF6) is a highly potent greenhouse gas with a Global Warming Potential (GWP) of 24,700 over 100 years and is globally mainly used as an electrical insulator in switchgear. Several measurement networks have tracked SF6 for many years and their European data reveal significant emissions in southern Germany. This study focuses on German SF6 emissions (2020–2023), using atmospheric measurements from 22 European sites, offering high spatial and temporal resolution for robust emission assessments. While German UNFCCC inventory bottom-up emission estimates report a major source of SF6 through the disposal of soundproof windows, the spatial distribution of German SF6 emissions derived on top-down inversion techniques (InTEM and Flexinvert+) reveals a different picture: The continuous pattern of high emissions from a particular region is responsible for one-third of total SF6 emissions in Germany. Despite this, total German SF6 emissions have decreased from 112 ± 26 t in 2020 to 89 ± 15 t in 2023 (InTEM), with estimates from all methods (both bottom-up and top-down) showing similar trends. Our findings suggest that the emissions from soundproof windows are overestimated, while industrial sources - particularly from SF6 production and recycling in the focus region - are likely underestimated.
2025
Ensuring data quality, completeness, and interoperability is crucial for progressing safety research, Safe-and-Sustainable-by-Design approaches, and regulatory approval of nanoscale and advanced materials. While the FAIR (Findable, Accessible, Interoperable, and Re-usable) principles aim to promote data re-use, they do not address data quality, essential for data re-use for advancing sustainable and safe innovation. Effective quality assurance procedures require (meta)data to conform to community-agreed standards. Nanosafety data offer a key reference point for developing best practices in data management for advanced materials, as their large-scale generation coincided with the emergence of dedicated data quality criteria and concepts such as FAIR data. This work highlights frameworks, methodologies, and tools that address the challenges associated with the multidisciplinary nature of nanomaterial safety data. Existing approaches to evaluating the reliability, relevance, and completeness of data are considered in light of their potential for integration into harmonized standards and adaptation to advance material requirements. The goal here is to emphasize the importance of automated tools to reduce manual labor in making (meta)data FAIR, enabling trusted data re-use and fostering safer, more sustainable innovation of advanced materials. Awareness and prioritization of these challenges are critical for building robust data infrastructures.
2025
Surveys in Norwegian schools showed that some students experienced health problems, such as headaches or concentration issues which have been linked to indoor environment quality (IEQ). This research investigates the relationship between measured IEQ and students’ perceived IEQ as user-feedback in one lower secondary school. This study explores the factors contributing to the connection with certain parameters such as carbon dioxide (CO2), volatile organic compounds (VOC), and temperature levels with perceived IEQ. Despite achieving good IEQ levels according to standards, there is a notable discrepancy between measured IEQ and how students perceive the air quality. Two classrooms served by a demand-controlled ventilation system were monitored with IEQ measurement sensors and online questionnaires were given individually to students in each classroom. This enables to provide real-time students’ perception of indoor air and room temperature quality. Measurement results showed IEQ are of good quality, but students’ responses on perceived IEQ vary and showed over 25% are dissatisfied, indicating mixed feelings and dissatisfaction about perceived IEQ. Future research should focus on refining ventilation systems to bridge the gap between measured and perceived IEQ.
2025
Circulating MicroRNAs in Cord Blood to Predict Attention-Deficit/Hyperactivity Disorder Diagnosis
Background
There are large knowledge gaps in the etiology of attention-deficit/hyperactivity disorder (ADHD), and although it is a prevalent and highly heritable neurodevelopmental disorder, diagnosis can be challenging. We aimed to assess the association of circulating blood plasma microRNAs (miRNAs) at birth with ADHD for use as biomarker candidates and build an miRNA-based prediction model.
Methods
Our study population consisted of 206 children with ADHD (33.0% female), 207 control children (33.8% female), and their parents from the MoBa (Norwegian Mother, Father, and Child Cohort Study). Expression levels of 51 selected miRNAs in plasma from children’s cord blood at birth and from both parents during early pregnancy were quantified by quantitative polymerase chain reaction and tested for association with children’s ADHD diagnosis and ADHD symptom scores based on ratings by parents.
Results
Seven miRNAs were differentially expressed at birth in children with ADHD and control children (false discovery rate < .05), and 31 had a statistically significant linear relationship with parent-rated ADHD symptom score at 8 years. A 19-miRNA ADHD prediction model achieved good discrimination in the test population (area under the receiver operating curve = 0.959, accuracy = 0.893). Functional analysis for the 19-miRNA prediction set revealed involvement in several highly relevant pathways, e.g., dopaminergic synapse, circadian rhythm, and axon guidance. We also found that parental miRNA expression levels significantly associated with children’s ADHD diagnoses and/or ADHD symptoms scores.
Conclusions
We showed that expression levels of circulating miRNAs at birth may be used to predict increased risk of ADHD diagnosis, and our 19-miRNA set should be included in future efforts to develop a biomarker panel.
2025
Building-related symptoms in school environment: Predictability using machine learning approach
Building-related symptoms (BRS) are commonly experienced by students in schools and are potentially affecting academic performance and health. Even though indoor environment quality (IEQ) measurements indicated fair conditions, students still perceived discomfort that led to symptoms, highlighting the necessity of collecting user-feedback about IEQ-complaints. This study aimed to predict and understand the prevalence of BRS (headache, tiredness, cough, dry eyes-hands) experienced by students in classrooms using machine-learning (ML) approach based on measurement data, building factors, and prevalence of IEQ-complaints. We collected measurement data (from indoor and outdoor climate), building factors, and user-feedback by students via online-platform across three sampled classrooms each campaign during three consecutive school semesters. Significant input variables for ML were pre-selected using statistical tests. ML models were evaluated based on accuracy metrics and SHAP analysis for input interpretation. Models using measurement data alone performed poorly (testing R² <50 %) to predict prevalence of BRS, whereas adding building factors and prevalence of IEQ-complaints increased accuracy (R² up to 95 %) of prediction of BRS with lower RMSE. In addition, interpretation from SHAP analysis showed IEQ-complaints especially related with indoor air quality (e.g., heavy air, dust & dirt, and dry air) as significant contributors for predicting prevalence of BRS. We conclude that the framework of combining objective measurements with occupant-reported complaints can be reliable, interpretable predictions of symptom prevalence. This study is limited by single-school setting, health confounders, and symptoms verification. Future research may contribute to exploring wider set of input variables, applicability, and variation of complaints preference.
2025