Many iron meteorites sample the metallic cores of planetesimals, providing a unique opportunity to study early Solar System planetary processes. Modeling the chemical evolution during crystallization of these iron meteorite groups provides evidence that both S and P were present as light elements in planetesimal cores. As fractional crystallization of these cores proceeded, the concentrations of both S and P in the residual metallic liquid in the core increased, and eventually, liquid immiscibility in the Fe-Ni-S-P system would be expected to be encountered in many cases. For example, the IIAB and IIG irons may be related in this way. Considerable previous experimental work has determined element partitioning behaviors between solid metal and liquid metal as a function of the S and P concentrations of the metallic liquid, but few experimental studies have examined trace element partitioning behavior after the onset of liquid immiscibility. In this work, we present new experimental results that determine element partitioning behavior between two coexisting immiscible metallic liquids in the Fe-Ni-S-P system, as may be experienced during the later stages of the evolution of planetesimal cores.

The reliable operation of transformers is critical to power system stability, and early fault detection in transformers can significantly reduce the risk of catastrophic failures. Dissolved gas analysis (DGA) is a commonly used method to assess transformer health, as gas concentrations correlate with fault conditions inside the transformer. Methane (CH4), a gas indicative of thermal faults, serves as a key parameter in detecting early signs of fault. This study proposes a linear epsilon support vector regression (SVR) model fine-tuned with Grid Search for predicting methane concentration based on transformer temperature, oil temperature, winding resistance, load factor, service life, and transformer age using grey relational analysis. By predicting methane concentration in ppm, this model supports proactive fault management, optimizing maintenance schedules, and effectively reducing transformer downtime. Comparative results suggest that the proposed SVR model is an effective baseline for improving dissolved gas analysis accuracy, contributing to enhanced transformer diagnostics and preventive maintenance strategies.

As conversational AI systems such as ChatGPT have become more advanced, researchers are exploring ways to use them in education. However, we need effective ways to evaluate these systems before allowing them to help teach students. This study proposes a detailed framework for testing conversational AI across three important criteria as follow. First, specialized benchmarks that measure skills include giving clear explanations, adapting to context during long dialogues, and maintaining a consistent teaching personality. Second, adaptive standards check whether the systems meet the ethical requirements of privacy, fairness, and transparency. These standards are regularly updated to match societal expectations. Lastly, evaluations were conducted from three perspectives: technical accuracy on test datasets, performance during simulations with groups of virtual students, and feedback from real students and teachers using the system. This framework provides a robust methodology for identifying strengths and weaknesses of conversational AI before its deployment in schools. It emphasizes assessments tailored to the critical qualities of dialogic intelligence, user-centric metrics capturing real-world impact, and ethical alignment through participatory design. Responsible innovation by AI assistants requires evidence that they can enhance accessible, engaging, and personalized education without disrupting teaching effectiveness or student agency.

In recent years, a lot of papers have been made investigating single case studies of Lithosphere-Atmosphere-Ionosphere-Coupling before earthquake occurrence by multiparametric researchers. So, a question naturally arises: Do we need to investigate another large earthquake?The answer is not trivial. Statistical studies are necessary and fundamental to support this research, as the occurrence of anomalies before an earthquake doesn’t prove that the anomalies are induced by the preparation of the earthquake. Contrariwise, by investigating statistically the time before the earthquake, it’s possible to compare if the rate of occurrence of specific anomalies is larger than the same rate in a non-seismic period, assessing the statistical significance of the pre-earthquake phenomenon. Anyway, when a lot of earthquakes are investigated at the same time, the details of the patterns of anomalies in the geo-layers are lost. For such a reason, the investigation of single case studies is precious if conducted in parallel with statistical investigations.In this framework, here, the investigation of a new earthquake is proposed: Mw = 7.2 14 August 2021 Haiti earthquake. The six months that preceded the Haiti earthquake are explored using data from the lithosphere, atmosphere, and ionosphere (among them with China's seismo-electromagnetic satellite and swarm satellites). Several anomalies have been depicted by different techniques. In particular, it’s very interesting to underline that all the geo-layers showed anomalies about two months before the seismic event: seismic accumulation of stress increased its slope, several atmospheric quantities underlined abnormal atmospheric conditions, and CSES-01 Ne depicted two consecutive days of ionospheric electron density in the six months preceding the earthquake. A final comparison with previously investigated earthquakes is provided trying to discuss the analogies and differences in the various events. In particular, the comparison with Mw = 7.7 Jamaica earthquake is crucial for the same availability of dataset, close location, and similar tectonic settings. The comparison could start to depict some reasons for the different patterns as the land or sea location of the epicenter.

This study tackles the limitations of traditional distance metrics in machine learning, particularly their inadequacy in managing missing data and varying feature importance. We introduce a novel custom distance metric that offers a more nuanced and robust measure of similarity, addressing the complexities and imperfections of real-world data. Our extensive evaluation of the CustomKNN algorithm, leveraging this custom distance metric, reveals exceptional performance and robustness across diverse datasets, consistently outperforming traditional metrics in various classification tasks. The custom distance metric shows superior robustness to missing data and scalability, despite its higher computational complexity and memory requirements. Our findings highlight the versatility and effectiveness of the CustomKNN algorithm, establishing it as a formidable option for classification tasks. This research advances the development of more adaptable and context-aware distance metrics, significantly enhancing the accuracy and reliability of machine learning models in practical applications.

The recognition of data as a natural resource has made headlines in the new era of industrialization. Companies now leverage this resource to enhance their services and products, often promising optimal outcomes. While data has always held significant value, recent advancements in AI and ML frameworks have brought this fact to the forefront. However, it was not until major scandals involving large corporations came to light that the critical issue of privacy was remembered. Consequently, the intersection of data, AI and ML frameworks, and privacy has emerged as a new area of research. Although numerous works have reviewed the development of various data protection techniques, it seems that most of them address the subject from a single perspective or attribute the entire concept of data privacy to a specific technique. This review aims to present an overarching view of the topic. It offers a systematic guideline that establishes a proper connection among the three elements: data, AI & ML frameworks, and privacy. The paper delves into each element from both an abstract and concrete standpoint, presenting the latest techniques to address data privacy concerns, including numerous lab simulations. It also recommends tools and resources for further study. Ultimately, it wraps up the central topic by outlining the challenges and prospective future research directions. Keywords: Data Privacy; AI & ML; Differential Privacy; Federated Learning;

The scientific debate on the existence of possible precursors before large earthquakes has been wide and brought to not univocal conclusions. One observable alone is clearly insufficient to provide indications on the criticality of the seismic fault status. In particular, it’s fundamental to investigate how critical is the accumulation of stress on a seismogenic fault and how we can predict an incoming earthquake. Considering that the Earth is a complex system in which the geo-layers (lithosphere, atmosphere and ionosphere) interact, it seems possible to understand better the Earth’s system if we study it as a whole. This requires an integration of observations and data from several sources and instruments, from ground data to remote sensing satellites. Despite the technical obstacles, several works have been done in this direction in the last years, and they seem to confirm firstly that the geo-layers show alterations (i.e., anomalies) before earthquake occurrence; secondly, it seems that several mechanisms of coupling are possible. This last point is the present research topic, as further clarifications are required.In the presentation, several examples of couplings before the earthquakes will be shown; for example, the recent investigation of the Lushan (China) 2013 earthquake, which shows three possible couplings by different geophysical mechanisms at very different times before the earthquake (130 days, 45 days, and 2 weeks).For the more recent seismic event, it is possible to investigate the ionosphere layer with the precious data from China Seismo Electromagnetic Satellite (CSES-01), by its instruments such as Langmuir probes and magnetometers. The integration of CSES and ESA Swarm data allows us to define better the ionospheric disturbances possibly related to earthquake occurrence. These cases include Mw = 7.5 Indonesia 2018, Mw = 7.6 Papua New Guinea 2019, Mw = 7.7 Jamaica 2020 or Mw = 7.8 Turkey 2023 earthquakes.Finally, we are still far from predicting earthquakes but several steps toward a better understanding of the phenomena eventually associated with seismic events seem to be not far.  

Water levels in lakes is a determinant of water availability for optimum water resources planning and ecosystem functioning. This study examines the surface water fluctuations in Kainji and Shiroro Lakes between 2008 and 2022 (14 years) using satellite altimetry. The purpose is to monitor the effect of precipitation, as an indicator of climatic variability on the water levels in the lakes considering their use for hydropower generation and flood control .From the observed values, there was no significant difference in water levels with the highest recorded at Kainji Lake as 138.84m and the lowest as 125.77m, however, there was significant difference in the  area coverage with the highest value as 1311.16km2  and lowest value as 761.15km2.Same trend was noticed for Shiroro Lake where the highest water level was 381.92m and lowest as 356.97m; there was a marked difference in area coverage with the highest area being 361.61km2 and lowest 104.36km2 . A further analysis was done to assess the spatiotemporal trend of the rainfall at the Kainji and Shiroro Lakes from rainfall data derived from TAMSAT. Using Mann Kendall (MK) and Spearman’s rho tests for the 14-year period, trend analysis for the Kainji Lake showed a slight positive trend for precipitation at a 5% significance level; with mean annual rainfall of 1137. 7mm.At the Shiroro Lake, there was no positive trend at a 5% significance level. The mean annual rainfall recorded was 1318.6m. Key Words: Precipitation, Significance level, Mann Kendall, Rainfall, Trend

The Tropical West Pacific (TWP) is the key entry point of air into the stratosphere during Northern Hemispheric winter. Our previous work showed how transport to the TWP troposphere is modulated by the movement of the Intertropical Convergence Zone, allowing transport of polluted air masses from tropical Asia during spring (Müller et al., 2023a, Müller et al. 2023b). Here, we present insights on air mass transport to the upper troposphere/lower stratosphere (UTLS) in this region using the ozone and relative humidity time series (2016-2023) from the Palau Atmospheric Observatory (PAO), located in the center of the tropical warm pool (7°N, 134°E). We further give an overview of all observations performed at the PAO during the ACCLIP summer campaign 2022, including ozone, CFH and COBALD sonde and lidar measurements.In analogy to Müller et al. 2023b, Lagrangian backward trajectories calculated are used to identify transport from remote locations. The time series already allows some interpretation regarding interannual variability mainly to the ENSO cycle. Here, we show anomalies during the 2016 strong El Nino and potentially the El Nino in the upcoming winter 2023. The PAO proves to be an excellent site to study UTLS composition and dynamics in the TWP region.

The geostationary Himawari-8 satellite offers a unique opportunity to monitor sub-daily thermal dynamics over Asia and Oceania, and several operational land surface temperature (LST) retrieval algorithms have been developed for this purpose. However, studies have reported inconsistency between LST data obtained from geostationary and polar-orbiting platforms, particularly for daytime LST, which usually shows directional artefacts and can be strongly impacted by viewing and illumination geometries and shadowing effects. To overcome this challenge, Solar Zenith Angle (SZA) serves as an ideal physical variable to quantify systematic differences between platforms. Here we presented an SZA-based Calibration (SZAC) method to operationally calibrate the daytime component of a split-window retrieved Himawari-8 LST (referred to here as the baseline). SZAC describes the spatial heterogeneity and magnitude of diurnal LST discrepancies from different products. The SZAC coefficient was spatiotemporally optimised against highest-quality assured (error < 1 K) pixels from the MODerate-resolution Imaging Spectroradiometer (MODIS) daytime LST between 01/Jan/2016 and 31/Dec/2020. We evaluated the calibrated LST data, referred to as the Australian National University LST with SZAC (ANUSZAC), against MODIS LST and the Visible Infrared Imaging Radiometer Suite (VIIRS) LST, as well as in-situ LST from the OzFlux network. Two peer Himawari-8 LST products from Chiba University and the Copernicus Global Land Service were also collected for comparisons. The median daytime bias of ANUSZAC LST against Terra-MODIS LST, Aqua-MODIS LST and VIIRS LST was 1.52 K, 0.98 K and -0.63 K, respectively, which demonstrated improved performance compared to baseline (5.37 K, 4.85 K and 3.02 K, respectively) and Chiba LST (3.71 K, 2.90 K and 1.07 K, respectively). All four Himawari-8 LST products showed comparable performance of unbiased root mean squared error (ubRMSE), ranging from 2.47 to 3.07 K, compared to LST from polar-orbiting platforms. In the evaluation against in-situ LST, the overall mean values of bias (ubRMSE) of baseline, Chiba, Copernicus and ANUSZAC LST during daytime were 4.23 K (3.74 K), 2.16 K (3.62 K), 1.73 K (3.31 K) and 1.41 K (3.24 K), respectively, based on 171,289 hourly samples from 20 OzFlux sites across Australia between 01/Jan/2016 and 31/Dec/2020. In summary, the SZAC method offers a promising approach to enhance the reliability of geostationary LST retrievals by incorporating the spatiotemporal characteristics observed by accurate polar-orbiting LST data. Furthermore, it is possible to extend SZAC for LST estimation by using data acquired by geostationary satellites in other regions, e.g., Europe, Africa and Americas, as this could improve our understanding of the error characteristics of overlapped geostationary imageries, allowing for targeted refinements and calibrations to further enhance applicability.

The increased occurrence of catastrophic events caused by climate change-induced mass movements affects human welfare and results in huge economic losses, making the development of early warning systems (EWSs) crucial to everyday decisions of local governments and communities. The state-of-the-art (SOTA) EWSs can only provide information on the level of direct danger of mass movement that each village is exposed to. This is the result of highly sensitive and coarse classifications and aggregated predictions at regional level, potentially leading to a poor perception of risk and inadequate local management plans. Also, this does not account for the indirect effects of mass movements on the local communities at social and economic levels, e.g., a town that can be cut out of the transportation and communication systems because of landslides blocking the roads around it is currently not considered by SOTA EWSs.To overcome this issue, we developed a novel machine learning scheme that investigates the environmental (hydrological and geological) characteristics of mass movements and the connectivity information of formal settlements and road network data in terms of graph structures. In particular, we study the interaction of the probabilistic mass movement susceptibility (derived from the environmental properties by means of a supervised ensemble graph neural network) on the graph representing the road network connecting the formal settlements. As a result, we derive for each formal settlement a probability of being indirectly affected by mass movements (e.g., the probability to be isolated as a result of mass movements affecting their surroundings) by graph spectral clustering.We tested this architecture (named Intergraph) on the Norwegian territory, taking advantage of over 68,000 incidents of reported mass movements since 1957. Our approach achieved an overall performance of 86.25% with the 2020 Gjerdrum quick clay incident as a demonstrated case study. With the intensifying effects of climate change, our study has opened an opportunity to develop solutions for adaptation and mitigation through a new holistic graphical perspective to assess various large-scale geospatial datasets of risk elements such as exposure, vulnerability, and hazard.

The Columbia root-knot nematode (Meloidogyne chitwoodi) is a destructive soil borne pest that can cause serious economic damage to potato tubers within infected, unfumigated fields. There are very few known sources of genetic resistance to root-knot nematodes and no released potato cultivars exhibit this trait. Literature indicates that nematode resistance introgressed from Solanum bulbocastanum is dominantly inherited across many genetic backgrounds. We generated a 32 family half-diallel progeny test population utilizing 3 root-knot nematode resistant clones (female) and 11 clones (male) with russet skin type (1,600 clones, between 25 – 60 clones per family). In 2023, 1,200 progeny were evaluated relative to 6 control varieties planted at high replication at the Washington State University Experiment Station in Othello, WA. A scale, RGB-D imaging conveyor, and index scoring system was used to record: total yield, the number of tubers per plant, tuber size distribution, aspect ratio, skin color, starch content, and defect severity from all samples within this population. The distribution of phenotypic values observed from the progeny suggest that choice of parent is a highly significant factor that influences almost all traits evaluated. The phenotyping strategy utilized by our team is inexpensive, expandable, and flexible enough to be adopted by many different types of vegetable breeding programs. Data from this experiment is being used to develop potato tuber defect classification models and assess the utility of adopting genomic selection within our potato breeding program.

Data science is an interdisciplinary subject that uses scientific methodologies, data mining techniques, machine-learning algorithms, and large amounts of data to extract information and insights. The article presents an overview of the current state and future possibilities of Data Science in a variety of sectors, explores the benefits, outlines the frameworks and methodologies employed, explains the current obstacles, and offers feasible solutions.

Light levels change throughout the day, are affected by climate and weather, and are filtered by the local environment. Switching between low and high levels of light over varying periods of time experienced by an organism in its environment shapes the tempo and mode of its light detection system. Plants must respond to dynamic environmental conditions and thus switch between efficient photosynthesis and photoprotection. Receptors on the plasma membrane perceive extracellular signals, such as photosynthetically-fixed sugars, are coupled to cytoplasmic G proteins to transduce information to cytoplasmic proteins and to amplify that signal to bring about changes like photosynthetic efficiency in both short (e.g. enzymatic reactions) and long (e.g. plant development) time scales. While G proteins have been shown to be important in regulating various aspects of stomata and photosynthesis, their role has yet to be fully understood. A regulator of G signaling (RGS) has been shown to sense sugars fixed in photosynthesis. Thus, we hypothesize that RGS mediates responses to dynamic light. The sequenced genomes within the grass family are the only genomes throughout Plantae known to lack RGS. By contrast, Setaria retains the RGS gene. Thus, the RGS gene from Setaria was expressed in rice to better understand the function of RGS. In this study, multiple transgenic events were grown to investigate their phenotypic response. We identified lines with altered stomatal patterning and rates of stomatal closure in response to changing light levels that will be used in future experiments.

A document by Karlyn Harrod. Click on the document to view its contents.

Quantifying 3D phenotypic traits for plant shoots and roots is essential to monitor and evaluate plant growth and development. Multi-view stereo (MVS) is a low-cost and widely used photogrammetry method to build 3D point clouds in many agricultural applications. However, it is challenging to adopt MVS directly to obtain complete 3D structures of fine roots for plants such as soybeans. To address this problem, we propose a data processing pipeline incorporating super-resolution (SR) and 3D Gaussian Splatting (GS) to enhance the resolution of 3D root reconstruction, aiming to recover a highly detailed 3D root structure. To this end, first, multi-view images of a soybean root are collected using an RGB camera; second, SR is used to optimize the resolution of the images; third, the processed images are fed to the algorithm structure from motion to obtain a point cloud; and then, 3D GS is applied to enhance the implicit 3D surface reconstruction; finally, perceptual similarity and peak signal-to-noise ratio (PSNR) are used to evaluate the output quality. The method is expected to obtain a high-fidelity 3D reconstruction of plant roots for soybeans and other crops, assisting in the extraction of comprehensive phenotypic traits to accelerate the selection of new varieties for plant breeding.

Hydrological observation networks are inadequate and declining, hindering advancements in hydrology, particularly in hydrological budgeting, where interception loss remains unmeasured, and a subjective loss figure is assumed. Measuring net rainfall reaching the ground surface is crucial for precise water balance estimates. This involves measuring throughfall & stemflow fractions of precipitation on the canopy. Current methods use collector channels and static storage systems requiring periodic visits for measurements, which becomes extremely difficult and unsafe in inaccessible forests, leading to data losses. For time-resolved data, an array of standard rain gauges is placed under the canopy, or water is collected by troughs for throughfall (and collars for stemflow measurement), draining into a central pipe and directed into a tipping bucket flow gauge. Commercial flow gauges are expensive for use in throughfall & stemflow measurement, since, multiple instruments are required in each study location to capture the variability of the canopy. To address this, we designed and fabricated an open-source, low-cost Tipping Bucket Flow Gauge to automatically monitor flow rates of throughfall & stemflow. It is designed to have a larger adjustable tipping resolution (10 ml – 200 ml). The open-source Arduino-based data logger automatically collects time-resolved data and is powered using solar energy, ensuring remote functionality even in harsh environments. A modular electronics approach was followed for designing the datalogger, facilitating rapid prototyping, easy repair, and upgrades, enabling someone with even an introductory knowledge in Arduino to implement the design. Almost 75% of the instrument is 3D printed and can be fabricated using any standard desktop FDM 3D printer and assembled by hand. The instrument showed 86% accuracy in preliminary testing at a calibrated tipping resolution of 120ml. The cost of prototyping was 100$ - 120$, thus proving cost-effective for accurate hydrological budgeting as compared to the cost of the nearest available commercial solution (~1800$). Through our research and product design, we intend to reduce the barrier of entry and simplify the steep learning curve faced in developing hydrological instrumentation.

Advances in quantitative genetics and high-throughput pipelines have allowed for rapid identification of genomic markers associated with changes in phenotype. However, linking those markers to causal genes is still difficult, as many genes may be linked to one marker. We aimed to improve candidate gene selection by creating a new method that identifies conserved genes underlying GWAS loci in multiple species. So far, we have tested this method in two different experiments: 1) using GWAS from Arabidopsis, soybean, rice, maize, and sorghum measuring 19 elemental uptake (ionomic) traits and 2) GWAS from Arabidopsis, soybean, rice, and maize measuring seed weight traits. We identified 14,336 candidates in the ionomics GWAS comparison. The most likely candidates belonged to ortholog groups linked to GWAS loci in all five species for their given trait according to calculations using random permutations of the data. For the seed weight GWAS comparison, we identified 192 candidates, and again, the most likely candidates belonged to ortholog groups linked to GWAS loci in all species in the comparison. Focusing on these most likely candidate genes from Arabidopsis, we obtained T-DNA lines with mutant alleles for each candidate gene and performed a high-throughput phenotyping screen utilizing ICP-MS for ionomics and the image analysis software PlantCV for seed weight. Preliminary results show 59 ionomic candidates and 9 seed weight candidates have one line with confirmed phenotypes. We plan to further verify these preliminary confirmations by obtaining and screening additional T-DNA lines with different alleles for each candidate gene.