High-resolution PM2.5 forecasts are increasingly produced with machine-learning models, yet practical guidance on how often these models should be retrained and validated remains limited. This study quantifies the impact of retraining frequency and bias correction on air-quality prediction skill across multiple cities with contrasting emission sources and meteorological regimes, using The Goddard Earth Observing System composition forecast (GEOS-CF) fused with in-situ observations. Site-specific models are trained on at least two years of hourly data, with bias correction and alternative update schedules (6-18-month baselines and 6-12-month retraining cycles) evaluated using RMSE, R2, and SHAP-based feature importance. Bias-corrected models consistently improves GEOS-CF forecasts by more than 107% in R2 and reduces RMSE by over 75%, with annual retraining providing the largest gains (13% increase in R2, 12% reduction in RMSE) relative to more frequent updates. SHAP analysis shows that the dominant predictors and their relative importance vary by city, with combinations of boundary-layer height, aerosol optical depth, humidity, wind, and nitrogen oxides driving PM2.5 levels, demonstrating that a single global pre-trained model is inadequate and that locally tuned models are required. Together, these results define minimum data requirements, preferred retraining intervals, and the need for site-specific bias-corrected models, offering concrete design rules for operational PM2.5 forecasting systems.
Many rivers and streams are affected by chemical pollution, yet current chemical monitoring methods are limited for technical and economic reasons. Biomonitoring has been increasingly used as a means of assessing the impacts of chemical pollution and indirectly monitoring river contamination. To provide an overview of the wide variety of biomonitoring approaches, we conducted a systematic review of the literature investigating the relationship between chemical pollution and biological responses. We distinguished five main approaches for the biomonitoring of chemical pollution: (i) monitoring of native communities; (ii) laboratory bioassays; (iii) in situ bioassays; (iv) mesocosms; and (v) monitoring of wild populations. Although each of them covers a wide range of methods and endpoints, we have highlighted their main advantages and limitations. Because native communities are exposed to a wide range of stressors, isolating the effects of chemical pollution alone is often limited. Most of the existing community indices cannot depict the full extent of the impact of pollutants on communities, but rather provide information on either general degradation of water and/or sediment quality. Effect-based methods (EBMs), including ecotoxicological bioassays and biomarkers, can better isolate the effects of pollution, and, to some extent, of specific types of pollutants. The experimental design of EBMs must be adapted to the research question and the context of the study, so that the test organisms, exposure scenarios and endpoints accurately reflect the contamination. In this context, a comparison with theoretically non-stressful situations with either a dilution series of the exposure solution in the laboratory or a comparison of laboratory and field treatment is relevant. The main difficulties encountered in the approaches investigated in the present review are the comparability of sampling strategies, non-linear concentration–response relationships, extrapolation from laboratory to field exposure, the highly variable sensitivity of organisms and the geographical specificities. Overall, a combination of different EBMs can integrate the effects of exposure to specific contaminants at both spatial and temporal scales while accounting for confounding factors. The establishment of thresholds and guidelines would facilitate the integration of EBMs into regular monitoring programmes. This in turn will greatly facilitate the assessment of chemical impairment.
BackgroundFor use in cosmetic products, 32 substances have been authorised in the EU as UV-filters as of August 2022 (Annex VI of the EU Cosmetics Regulation (EC) No 1223/2009). Although these substances can enter the aquatic environment directly through bathing, authorisation for use in cosmetic products does not require a specific environmental risk assessment, whereas risks to human health are specifically assessed. Furthermore, no environmental quality standards have been proposed at EU level. For the current review, data on freshwater ecotoxicity, physicochemical properties, environmental fate and measured exposure were retrieved from ECHA registration information and public databases and assessed for reliability and relevance. Environmental quality standards (EQS) were proposed based on the EU technical guidance for EQS for the Water Framework Directive (WFD).ResultsFrom a WFD perspective, there were significant or complete gaps for acute and chronic effects data, sediment organisms and rarely tested organisms at the time of data retrieval, as well as gaps for measured environmental concentrations in the assessed databases. The reliability of the reported data could not always be verified due to missing information. As expected from the heterogeneity of molecular structures, ecotoxicity data and environmental fate parameters vary considerably between the different substances. Our assessments indicate that the use of some of the substances in direct contact with surface water may pose a repeated risk to aquatic organisms.ConclusionsWe recommend a revision of the process for the inclusion of UV-filters in the EU Cosmetics Product Regulation, taking into account the potential environmental risks. A full assessment of all UV-filters using standard test organisms as well as rarely tested species for acute and chronic effects and bioaccumulation would require, among other things, a large number of fish tests. Therefore, new assessment methods (NAMs) are recommended, including the use of fish cell lines for acute toxicity testing and alternatives for the assessment of chronic toxicity.
The growing use of algorithmic and biometric technologies in border control is part of a larger trend in global security governance that has significant legal and ethical implications for their effect on individual rights and procedural justice. As central features in the EU's shifting security regime, ETIAS and facial recognition technologies deploy algorithmic profiling and biometric risk assessment to screen visa-exempt third-country nationals. The research systematically examines the decision-making processes of ETIAS and the overall facial recognition system, demonstrating the interplay between algorithmic risk assessments and discretionary human discretion by national authorities. It contends that the algorithmic profiling lack of transparency, combined with sweeping national security exceptions, produces a procedural void, in which the right to reasoned decisions and effective remedies is compromised. Second, the use of interoperable databases and risk indicators puts core data protection principles into jeopardy, notably purpose limitation and the right to be forgotten. This paper also argues that ETIAS and the application of facial recognition technologies represent a larger trend toward “techno-regulatory assemblages” in EU governance, where technological infrastructures increasingly influence legal and administrative decisions. It critically assesses whether the human oversight mechanisms incorporated within ETIAS National Units are adequate to prevent the risks involved in automated decision-making, especially in the face of strict time pressures and security requirements. The study detects a latent paradox: though these systems aim to strengthen a “Security Union,” they might inadvertently lead to an “Insecurity Union” by undermining transparency, procedural protections, and citizen rights.
Abstract Islamic law does not specify a fixed marriage age. Rather, the walī can conclude a marriage for a minor at any age. To conclude their own marriage, prospective spouses must be of sound mind ( ʿāqil ) and pubescent ( bāliġ ), with a previously unmarried girl only being allowed to do so according to the Hanafi school. In an effort to curb early marriage, a practice that involves predominantly girls, legislatures in Muslim jurisdictions have introduced a fixed legal age of marriage and have increased state oversight over the authorisation of early marriages. At the same time, however, many countries continue to apply the Islamic concepts of physical and mental maturity. As a result, provisions based on Islamic law and Western legal concepts interact and, at times, exist in tension with one another. Subsequently, the flexibility of Islamic law has been preserved and early marriage continues to be legally valid. These marriages are concluded for a wide variety of reasons, encompassing economic, social and religious factors. Thus, legal reform is only one of many tools needed to meet the challenges posed by early marriage.
