62880 research outputs found
Sort by
Extended lead‐time geomagnetic storm forecasting with solar wind ensembles and machine learning
Full text linkGeomagnetic storms are large disruptions of the magnetosphere, which can impact satellites, communications systems, and power grids, causing significant technological and economic impacts. Current forecasting models utilize L1 satellite data, constraining lead time to a few hours, often insufficient for effective mitigation. We investigate how to extend the lead times of these forecasts with solar data. Associated spatial and propagation uncertainties of solar data are captured with a solar‐wind ensemble, of the computationally efficient one‐dimensional HUXt numerical model. The solar‐wind ensemble once propagated to Earth is processed through logistic regressions, weighting ensemble members by comparison with historical observed velocities, effectively filtering out high error ensemble members. Performance was evaluated across different storm intensities and lead times, demonstrating the models predictive capabilities in a variety of circumstances. Although not including transient phenomena such as Coronal Mass Ejections, our approach demonstrates strong predictive capability, achieving a Brier Skill Score relative to climatology (BSSclim) of 0.595 and a Receiver Operating Characteristic Area Under the Curve (ROC AUC) of 0.751 at 6‐hr lead time for storms defined as Hp30MAX ≥ 5 within a 24‐hr forecast window. Overall, these results highlight the strong potential of the coupled numerical model and machine learning framework to extend the forecast lead time for geomagnetic storms
Differential effects of agricultural expansion on wild bee taxonomic and functional diversity
No full textLos polinizadores, especialmente las abejas, están en declive global, amenazando la biodiversidad y la seguridad alimentaria. Si bien la agricultura intensiva es un factor principal, su impacto sobre la diversidad funcional de abejas—particularmente en la diversa región mediterránea—permanece poco estudiado. Comprender cómo la pérdida de hábitat natural en paisajes agrícolas afecta la diversidad funcional es crucial para desarrollar un manejo del territorio que preserve los servicios de polinización.
Evaluamos cómo la proporción de área natural afecta la diversidad taxonómica y funcional de abejas silvestres en huertos de manzano y cerezo en Chile central. Durante dos años, las abejas fueron muestreadas utilizando trampas de plato coloreadas en huertos categorizados según la vegetación natural circundante. La diversidad taxonómica fue analizada mediante índices de riqueza y composición, mientras que la diversidad funcional fue calculada a partir de rasgos morfológicos y de historia de vida, analizados separadamente para machos y hembras.
Encontramos que una mayor proporción de áreas naturales afectó positivamente la diversidad taxonómica y funcional de abejas silvestres en los huertos. La diversidad taxonómica de abejas aumentó significativamente con la proporción de área natural circundante, pero las respuestas de la diversidad funcional fueron contrastantes entre los rasgos de abejas macho y hembra. En particular, la homogeneización funcional (basada en RaoQ) se asoció significativamente con los rasgos funcionales de machos en huertos con baja proporción (%) de áreas naturales y, en algunos casos, la riqueza funcional (FRic) aumentó con una mayor proporción de área natural circundante. Encontramos que la diversidad taxonómica y funcional de abejas respondió de manera diferente en huertos de manzano y cerezo. Además, el análisis de las cuarto esquinas mostró algunas asociaciones sexo‐específicas y temporales entre rasgos y el porcentaje de áreas naturales.
Las áreas naturales parecen promover la diversidad taxonómica de polinizadores y la diversidad funcional de abejas de maneras distintas. Los enfoques basados en rasgos sexo‐específicos revelan asociaciones ecológicas diferenciales. Por lo tanto, sugerimos que futuros estudios incluyan la variabilidad intraespecífica de rasgos (como el dimorfismo sexual) parece relevante, y se requiere una comprensión más profunda de cómo los diferentes aspectos de la diversidad taxonómica y funcional de abejas responden a los filtros ambientales impuestos por las áreas naturales vecinas a los cultivos
Who taught you to torture? Punk and the Marquis de Sade
No full textThis is a short essay for an exhibition catalogue explaining how punk may be understood with regard the Marquis de Sad
Enhancing maritime situational awareness through multimodal fusion: insights from a real-world experiment
No full textEffective maritime border surveillance is crucial. Challenges we face include irregular migration, smuggling, oil spills and the need for rapid search and rescue. Various sensing technologies, including AIS, SAR, optical and infrared sensors, as well as UAV-mounted sensors, clearly enhance maritime awareness. However, integrating their diverse outputs remains complex. Feature-level multi-modal sensor fusion is a well-known methodology for robust detection and behavior analysis. However, most research relies on simulations or isolated sensors, which limits practical insights.
This study presents a controlled real-world experiment combining synchronized data from coastal ground sensors and UAV-mounted visual and infrared sensors. The recorded dataset enables the evaluation of feature-level fusion in authentic conditions. We enhance existing fusion frameworks with additional modules and assess them using operational metrics. This study contributes to our understanding of the efficacy of multi-modal fusion in complex maritime environments, while also highlighting the significant challenges involved in transitioning from simulations to controlled real-world sensor data
Bias adjustment and the question of usable climate information: methodological assumptions and value judgements
No full textStatistical bias adjustment has become a common practice to increase the relevance of climate model outputs for impact studies and other societal applications. However, the application of bias adjustment raises fundamental issues identified in the literature, calling into question the credibility of the adjusted climate information. In the attempt to address the usability gap of climate model output despite these unresolved issues, different approaches to bias adjustment have emerged—from applying a single consistent method across studies, selecting the most suitable method for a given use case, to employing an ensemble of bias adjustment methods. This paper examines how these approaches rest on both methodological assumptions and implicit value judgments about what constitutes usable climate information and for whom it is produced. Building on recent literature in the philosophy of science, we propose a framework for evaluating the usability of climate projections in the context of bias adjustment and apply this framework to evaluate the different approaches to bias adjustment. To evaluate the credibility of the adjusted climate information, the paper provides a detailed discussion of two key methodological assumptions underlying different approaches, the interpretation of performance differences of bias adjustment methods and changes to the climate model trend and ensemble through bias adjustment. Through this perspective, we aim to situate bias adjustment in the discussion around usable climate information and the production of climate services, while offering a practical discussion of assumptions for climate impact researchers and climate service practitioners working with bias adjustment methods
Calibrating probabilistic solar‐wind forecasts driven by the Wang‐Sheeley‐Arge model
Full text linkBy spatially perturbing coronal model output within a coupled coronal-heliospheric model we can generate probabilistic predictions of solar-wind speed. We apply these spatial perturbations to the Wang-Sheeley-Arge (WSA) model output to generate large sets of input conditions for the Heliospheric Upwind eXtrapolation with time dependence (HUXt) solar-wind model. The resulting ensemble forecasts at 1 AU contain useful information about likely outcomes and the method allows uncertainty to be better characterized. We tune the scales of perturbations to calibrate the probabilistic predictions. We use the rank histogram and reliability component of the Brier score to demonstrate how increasing levels of perturbation/variability generally improves the reliability of the WSA-HUXt ensemble distribution; the ability of the ensemble to capture the true likelihood of events based on observational frequencies. We use the resolution component of the Brier score to highlight how too large a perturbation harms the statistical resolution of the forecast; the ability of the model to meaningfully distinguish between events beyond a statistical observational baseline (like climatology). This adds a useful constraint on the maximum size of perturbation we should be applying. Additionally, we use continuous ranked probability score to demonstrate how a calibrated ensemble can improve a prediction system, reducing forecast error across all lead times. Finally we demonstrate that the calibrated ensemble provides value for an end-user through a Cost/Loss analysis. In refining this calibration procedure we provide optimal values of the perturbation parameters for use in the operational WSA-HUXt forecast
Temporal insights into afforestation: a 25-year study of woodland expansion
Full text linkDirect observation of transitions from farmlands to woodlands are rare, and habitat change dynamics are typically inferred indirectly from pseudo-chronosequence observations (space-for-time substitution). The aim of this study was to directly examine the transition of plant communities and soil properties as farmland developed into ex-arable planted woodland, in comparison to nearby ancient woodland. We collected plant identity and abundance records and topsoil cores for chemical analysis between 2001 and 2024. We additionally collected canopy openness data and soil cores for identification of soil fungal communities in 2024, to link plant communities and soil chemical properties with fungal communities and light availability. We modelled differences in species distribution across gradients of distance from ancient woodland. Our results showed that soil properties and plant communities in ex-arable and ancient woodlands remained significantly different 25 years after land- use change. Ex-arable woodland plots had higher pH, more available phosphorus, less total nitrogen, and less total carbon than ancient woodland plots. Changes in plant diversity were mostly associated with carbon to nitrogen ratio. The presence of ancient woodland indicator plant species was associated with higher light availability and proximity to ancient woodland edge. Ectomycorrhizal fungal communities were also significantly different in ex-arable and ancient woodlands, with higher fungal richness positively associated with soil pH and tree species richness. Our study suggests that woodland creation programmes should consider connectivity to ancient woodlands, time lags associated with plant and fungal succession, as well as the history of soil alteration practices, especially within woodland creation and compensatory schemes
Verification of AI–based environmental forecasting systems: what can we do, what do we need to do, and what are the challenges?
Full text linkSeveral institutions have released global medium–range meteorological forecasting models based on methods from machine learning, with training data provided by various reanalysis experiments. A proper and in-depth assessment of these models and the quality of their forecasts has yet to be carried out. Although in terms of simple and overall measures of skill such as mean square errors, AI-based forecasts clearly show very promising skill, we are just beginning to understand where and when these forecasts are useful and when they are not. Furthermore, while verification of meteorological forecasts has been subject to extensive (and still ongoing) research with a well established core methodology, it is not clear to what extent this methodology needs to be adapted or modified for AI–based models. Our paper aims to provide a vision on the verification of AI–based weather forecasts, identifying challenges, outlining important research questions, and laying the groundwork for a methodology to assess the quality of such forecasts
Generative AI and the documentary archive: creative opportunities and ethical abuses
No full textOne of the most innovative uses of AI in documentary filmmaking has been its deployment in animating the audio archive (Lees, 2023), for instance enabling in-vision interviews to be generated from twentieth century sound recordings. More recently, Generative AI has been deployed to manipulate archival photographs, however such uses by documentarists have led, in certain cases, to storms of controversy (What Jennifer Did, Jenny Popplewell 2024). This chapter examines the practices of documentary filmmakers using AI to alter or augment archival sources, and the ethical issues raised. It develops a theoretical approach to the documentary archive in the age of AI, arguing that Jamie Baron’s concept of the ‘archive effect’ (2014) can now be updated to describe a new ‘GenAI archive effect’, in which the spectator reflects on the realism of the fake, sometimes indistinguishable from actuality video