Modeling Efforts Research Articles

Machine Learning-Driven Scattering Efficiency Prediction in Passive Daytime Radiative Cooling

Passive daytime radiative cooling (PDRC) has emerged as a promising, electricity-free cooling approach that reflects sunlight while radiating heat through the atmospheric transparent window. However, the design and optimization of PDRC materials remain challenging, requiring significant time and resources for experimental and numerical modeling efforts. In this work, we developed a machine learning (ML)-driven approach to predict scattering efficiency in the wavelength of 0.3–2.5 μm, with the aim of eventually optimizing the microstructural design of PDRC materials. By employing ML models such as linear regression, neural networks, and random forests, we aimed to predict and optimize the scattering efficiency across different pore sizes and mixed-pore-size configurations. As a result, the random forest model demonstrated superior prediction performance with minimal error, effectively capturing complex, non-linear interactions between material features. We also leveraged data transformation techniques such as one-hot encoding for generative predictions in mixed-pore-size configurations. The presented ML-driven platform serves as a valuable open resource for PDRC researchers, facilitating the rapid and cost-effective optimization of PDRC materials and accelerating the development of sustainable cooling technologies.

The Quasi-Continuous Exhaust regime in JET

Abstract The quasi-continuous exhaust (QCE) regime is a regime that is naturally type-I ELM-free. It combines the high density at the plasma edge needed for power exhaust with the high normalised energy confinement typical for H-mode operation. In the QCE regime large-scale ELMs are avoided and high-frequent, low-amplitude filaments are present leading to the name-giving quasi-continuous edge transport of particles and energy.
This contribution reports that for the first time the QCE regime was successfully achieved in JET with a metal wall. More so, it was demonstrated in the recent JET deuterium-tritium campaign DTE3 that the regime is compatible with D-T operation.
Bringing the QCE regime to JET strongly benefited from the experimental and modelling efforts at the medium sized tokamaks ASDEX Upgrade and TCV. Using the physics picture developed from the ASDEX Upgrade experimental results, the route to the QCE regime in JET reported here is following closely the approach that was successful in ASDEX Upgrade. First, strong plasma shaping - large elongation and triangularity and the highly correlated closeness to double null - is developed. Second, sufficient fuelling to achieve high enough density at the pedestal foot, close to the separatrix, is applied. In addition, neon seeding proved to be very beneficial to avoid type-I ELMs when reducing the main ion fuelling.

Protecting Dynamically Obfuscated Scan Chain Architecture from DOSCrack with Trivium Pseudo-Random Number Generation

Design-for-test/debug (DfT/D) introduces scan chain testing to increase testability and fault coverage by inserting scan flip-flops. However, these scan chains are also known to be a liability for security primitives. In previous research, the dynamically obfuscated scan chain (DOSC) was introduced to protect logic-locking keys from scan-based attacks by obscuring test patterns and responses. In this paper, we present DOSCrack, an oracle-guided attack to de-obfuscate DOSC using symbolic execution and binary clustering, which significantly reduces the candidate seed space to a manageable quantity. Our symbolic execution engine employs scan mode simulation and satisfiability modulo theories (SMT) solvers to reduce the possible seed space, while obfuscation key clustering allows us to effectively rule out a group of seeds that share similarities. An integral component of our approach is the use of sequential equivalence checking (SEC), which aids in identifying distinct simulation patterns to differentiate between potential obfuscation keys. We experimentally applied our DOSCrack framework on four different sizes of DOSC benchmarks and compared their runtime and complexity. Finally, we propose a low-cost countermeasure to DOSCrack which incorporates a nonlinear feedback shift register (NLFSR) to increase the effort of symbolic execution modeling and serves as an effective defense against our DOSCrack framework. Our research effectively addresses a critical vulnerability in scan-chain obfuscation methodologies, offering insights into DfT/D and logic locking for both academic research and industrial applications. Our framework highlights the need to craft robust and adaptable defense mechanisms to counter evolving scan-based attacks.

Transonic Turbulence and Density Fluctuations in the Near-Sun Solar Wind

Abstract We use in situ measurements from the first 19 encounters of Parker Solar Probe and the most recent five encounters of Solar Orbiter to study the evolution of the turbulent sonic Mach number M t (the ratio of the amplitude of velocity fluctuations to the sound speed) with radial distance and its relationship to density fluctuations. We focus on the near-Sun region with radial distances ranging from about 11 to 80 R ⊙. Our results show that (1) the turbulent sonic Mach number M t gradually moves toward larger values as it approaches the Sun, until at least 11 R ⊙, where M t is much larger than the previously observed value of 0.1 at and above 0.3 au; (2) transonic turbulence with M t ∼ 1 is observed in situ for the first time and is found mostly near the Alfvén critical surface; (3) Alfvén Mach number of the bulk flow M A shows a strong correlation with the plasma beta, indicating that most of the observed sub-Alfvénic intervals correspond to a low-beta plasma; (4) the scaling relation between density fluctuations and M t gradually changes from a linear scaling at larger radial distances to a quadratic scaling at smaller radial distances; and (5) transonic turbulence is more compressible than subsonic turbulence, with enhanced density fluctuations and slightly flatter spectra than subsonic turbulence. A systematic understanding of compressible turbulence near the Sun is necessary for future solar wind modeling efforts.

Effect of non-spherical shape of dry powder aerosol drugs on their airway deposition distribution.

In the majority of aerosol drug deposition modelling efforts, the particles are approximated by regular spheres. However, microscope images acquired after drug formulation available in the open literature suggest that their shape is not regular in most cases. This work aimed to combine experimental measurements and numerical simulations to reveal the shape factors of the particles of commercialized aerosol drugs and the effect of non-sphericity on the lung deposition distribution of these drugs. Aerosol particles of Bretaris® Genuair®, Buventol® Easyhaler® and Trelegy® Ellipta® were collected on the stages of a Next Generation Impactor after their emission from the dry powder inhalers. The samples were analyzed using scanning electron microscopy techniques. Aspect ratios and dynamic shape factors characteristic of particles in the size fractions representing each impactor stage were determined. Computer modelling of their airway deposition was performed by both neglecting and considering their irregular shape. The results of particle size measurements revealed that there was no size-specific shape factor of the particles, as they could be well characterized by a global shape factor. Although there was some inter-drug difference in terms of shape factors, the values were quite close to unity (usually between 1.02 and 1.06), except for the fiber-shaped particles representing a minority in the composition of Trelegy® Ellipta®, which could be characterized by shape factors of around 1.6. The results of computer simulations of deposition distribution indicate that neglecting the irregular shape does not lead to a major distortion of the simulation results unless fiber-shaped particles are also present after the formulation.

Development and validation of a high-throughput qPCR platform for the detection of soil-transmitted helminth infections.

Historically, soil-transmitted helminth (STH) control and prevention strategies have relied on mass drug administration efforts targeting preschool and school-aged children. While these efforts have succeeded in reducing morbidity associated with STH infection, recent modeling efforts have suggested that expanding intervention to treatment of the entire community could achieve transmission interruption in some settings. Testing the feasibility of such an approach requires large-scale clinical trials, such as the DeWorm3 cluster randomized trial. In addition, accurate interpretation of trial outcomes requires diagnostic platforms capable of accurately determining infection prevalence (particularly as infection intensity is reduced) at large population scale and with significant throughput. Here, we describe the development and validation of such a high-throughput molecular testing platform. Through the development, selection, and validation of appropriate controls, we have successfully created and evaluated the performance of a testing platform capable of the semi-automated, high-throughput detection of four species of STH in human stool samples. Comparison of this platform with singleplex reference assays for the detection of these same pathogens has demonstrated comparable performance metrics, with index assay accuracy measuring at or above 99.5% and 98.1% for each target species at the level of the technical replicate and individual extraction respectively. Through the implementation of a rigorous validation program, we have developed a diagnostic platform capable of providing the necessary throughput and performance required to meet the needs of the DeWorm3 cluster randomized trial and other large-scale operational research efforts for STH. Resulting from the rigorous developmental approach taken, the platform we describe here provides the needed confidence in testing outcomes when utilized in conjunction with large-scale efforts such as the DeWorm3 trial. Additionally, the successful development of an evaluation and validation strategy provides a template for the creation of similar diagnostic platforms for other neglected tropical diseases.

Improving neuroendocrine tumor treatments with mathematical modeling: lessons from other endocrine cancers

Neuroendocrine tumors (NETs) occur sporadically or as part of rare endocrine tumor syndromes (RETS) such as Multiple Endocrine Neoplasia 1 and Von-Hippel Landau syndromes. Due to their relative rarity and lack of model systems, NETs and RETs are difficult to study, hindering advancements in therapeutic development. Causal, or mechanistic mathematical modeling, is widely deployed in disease areas such as breast and prostate cancers, aiding understanding of observations as well as streamlining in vitro and in vivo modeling efforts. Mathematical modeling, while not yet widely utilized in NETs research, offers an opportunity to accelerate NET research and therapy development. To illustrate this, we highlight examples of how mathematical modeling associated with more common endocrine cancers has been successfully used in the preclinical, translational, and clinical settings. We also provide a scope of the limited work that has been done in NETs and map how these techniques can be utilized in NET research to address specific outstanding challenges in the field. Finally, we include practical details such as hardware and data requirements, present advantages and disadvantages of various mathematical modeling approaches, and discuss challenges of using mathematical modeling. Through a cross-disciplinary approach, we believe that many currently difficult problems can be made more tractable by applying mathematical modeling and that the field of rare diseases in endocrine oncology is well-poised to take advantage of these techniques.

Potential serotype-specific effectiveness against IPD of pneumococcal conjugate vaccines V114 and PCV20 in children given a 2+1 dosing regimen

ABSTRACT Background Next generation, higher valency pneumococcal conjugate vaccines (PCVs) are assessed and licensed by comparing the immune response across serotypes shared with the PCVs that are standard of care for prevention of pneumococcal disease. Methods Using a previously qualified method we predicted the serotype-specific vaccine effectiveness (VE) against invasive pneumococcal disease of V114 and PCV20 for the serotypes shared with PCV13 in an EU, Russian, and Australian pediatric population that is recommended to receive a 2 + 1 dosing regimen. Results The estimated protective antibody concentrations ranged from 0.03 (serotype 23F) to 1.49 µg/mL (serotype 19F). Predicted VE values for V114 ranged from 79% (serotype 5) to 100% (serotype 23F). V114 had comparable effectiveness to PCV13 for all but one of shared serotypes, with predicted higher effectiveness (in V114) against serotype 3 (93% vs. 65%). Predicted VE values for PCV20 ranged from 47% (serotype 3) to 91% (serotype 14). PCV20 predicted VE was lower than PCV13’s for serotypes 4, 19F, 23F, 1, 3, 5, 6A, 7F, and 19A. Conclusions Predicted serotype-specific VE values suggest that, with a 2 + 1 dosing regimen, V114 will have greater effectiveness than PCV20 against PCV13 serotypes, particularly for the still-prevalent serotype 3. Real-world VE studies will ultimately provide clarity on the effectiveness of novel PCVs and support further confidence in and/or improvements to modeling efforts.

Physician Assistants in Clinical Endocrinology: Characteristics and Demographics.

Physician assistants (PAs) are employed in endocrinology, but little is known about their roles and activities. The study aimed to assess PAs' employment characteristics in endocrinology compared to those in all other specialties. This descriptive observational study used the 2022 National Commission on Certification of PAs (NCCPA) dataset. The study includes 117,748 board-certified PAs who indicated a clinical specialty in 2022. The characteristics of PAs in endocrinology were examined using descriptive statistics, including counts and percentages for categorical variables; means (with standard deviations), and medians (with interquartile ranges) for continuous variables. Bivariate analyses (ꭙ2 and Mann-Whitney U tests) were used to determine statistical differences between PAs practicing in endocrinology vs. PAs in all other specialties. This study found that as of 2022, 685 PAs reported practicing in endocrinology. PAs in endocrinology, compared to PAs in all other specialties (all p<0.001), were more likely to identify as female (82.0% vs. 69.6%), work in an office-based private practice (61.3% vs. 37.0%), and participate in telemedicine (70.8% vs. 40.1%). Conversely, PAs in endocrinology were less likely to work in a secondary position, saw slightly fewer patients weekly, and earned $10,000 less yearly than their PA colleagues in all other specialties. Examining the PA endocrinology workforce is essential due to the shortage of endocrinologists and the increased prevalence of diabetes as the U.S. population ages. Understanding where PAs in endocrinology are employed and their attributes could assist efforts in specialty modeling to address supply and demand projections.

Phantom Cooling Effects on Rotor Blade Surface Heat Flux in a Transonic Full Scale 1 + 1/2 Stage Rotating Turbine

Abstract An experimental and numerical investigation of phantom cooling effects on cooled and uncooled rotating high pressure turbine blades in a full scale 1 + 1/2 stage turbine test is presented. Objectives set to capture, separate, and quantify the effects of upstream vane film-cooling and leakage flows on downstream rotor blade surface heat flux. Multiple series of 1 + 1/2 stage rotating high pressure turbine tests were carried out in the Air Force Research Laboratory, Turbine Research Facility, at Wright-Patterson Air Force Base, Ohio. A non-proprietary research turbine was instrumented with high frequency double-sided thin film heat flux gauges custom made at AFRL. High bandwidth, time-resolved surface heat flux is measured on multiple film-cooled and non-film-cooled HPT rotor blades downstream of both film-cooled and non-film-cooled vane sectors. Upstream wake passing and heat flux is characterized on both rotor pressure and suction side surfaces, along with quantifying rotor phantom cooling effects from non-uniform first-stage vane film cooling and leakage flows. Fast response heat flux measurements quantify how rotor phantom cooling impacts the blade pressure side greatest. Measurements show upstream vane film-cooling alone can account for 50% of the rotor blade cooling effect in some instances, and even outweigh the rotor blade film cooling effect far from the blade showerhead holes. Added unsteady aero numerical simulation demonstrates how variations in inlet total temperature and incidence angle also contribute to the circumferentially non-uniform rotor heat flux. Results from this investigation aid modeling and design efforts in optimizing film cooling performance in real high-pressure turbine flow fields. Understanding the behavior of such non-uniform circumferential rotor phantom cooling effects can be critical to optimizing the efficiency, fuel consumption, range, and durability of advanced turbomachines.

Research on Vibration Comfort in Modular Construction

This paper focuses on the vibration problem of the first modular school construction project in Guangming District, Shenzhen. Field tests were conducted on modular classrooms under jumping scenarios to obtain vibration acceleration time–history curves, and the fundamental frequency of the modular classroom floors was determined using the Fast Fourier Transform (FFT). Subsequently, tests under walking and running scenarios were carried out to collect vibration acceleration time–history curves. By comparing the measured peak vibration accelerations with the limits specified in existing standards, potential vibration comfort issues in the modular classrooms were identified. A FES (finite element simulation) approach was employed to develop a comprehensive model of the modular school under investigation and to analyze its vibration comfort across the tested scenarios. This modeling effort served to validate the accuracy of the experimental measurements obtained from field testing. Finally, the vibration comfort of the modular classroom floors was further analyzed using finite element simulations. The results indicate that the modular classroom floors have significant vibration comfort issues.

Scale dependence of bird diversity in London

ContextUnderstanding drivers of biodiversity in cities can be mutually beneficial for ecosystems and people. Crowd-sourced bird observations provide an opportunity to assess how patterns of bird diversity change across observation scales and suggest driving processes.ObjectivesWe assessed the scale dependence of bird diversity within a 128 × 128 km extent over London’s urban–rural gradient to suggest scales at which key drivers may be operating.MethodsWe quantified scale variance of bird diversity across scales from 500 m to 64,000 m for three groups of species (All, Passeriformes, and Anseriformes and Charadriiformes combined). We estimated diversity by aggregating observations into a series of grids and computed comparable diversity estimates within each cell using interpolation and rarefaction. We calculated the variance explained by each scale for common diversity metrics.ResultsThe results show that bird diversity patterns around London vary by scale, and that the location of high variance shifts across the study area depending on both scale and species group. The variance of Passeriformes diversity gradually shifted from the urban core to the periphery, while variance of Anseriformes and Charadriiformes diversity occurred near water features.ConclusionsThe results suggest that the urban–rural gradient and location of water are two properties of the study extent around London influencing the scale dependance of bird diversity that could be used to ground scale considerations of further modeling efforts.

Sustainability language found in forest plans and its mathematical modeling potential

Over the last fifty years, management plans have become more descriptive with regard to the potential sustainability of forest systems, raising questions about the feasibility of implementing management activities and ensuring the sustainability of a wide variety of ecosystem services. To assess this issue, we conducted a survey among forest planning and operation research communities in the United States to understand their perceptions regarding the potential of a sample of sustainability statements currently used in national forest plans to be incorporated into optimization models or other mathematical operations. A total of 65 people responded to the survey, resulting in a response rate of 16.5%. Almost 70% of the respondents reported having over 30 years in their specified fields. The results revealed that only 3 out of 15 statements from the sample were deemed to have relatively mature or firm methodologies and data to enable inclusion in modern mathematical models for land use optimization. Of the three statements, two were related to economic sustainability, offering quantifiable information such as a non-declining flow of wood products over time and limits on the amount of timber harvested per decade. In contrast, sociocultural and, to some extent, ecological statements regarding sustainability were generally perceived to be more difficult to translate into mathematical modeling efforts. Particularly challenging were statements corresponding to sustaining natural or scenic characteristics of a forest. These findings may be attributed to various factors, including a lack of measurable indicators for sustainability and a potential lack of understanding about the modeling components and their interactions with planned management activities.

Synergistic image and point cloud processing of UAV data for urban flood modeling: point cloud smart thinning and curb mapping

Abstract. We propose an integrated approach for automatic point cloud thinning and curb mapping in Uncrewed Aerial Vehicle - Structure from Motion (UAV-SfM) point clouds to enhance hydrological modeling in flood-prone urban areas. UAV flights were conducted to generate an initial orthoimage, which was used to train a convolutional neural network (CNN) segmentation model. The trained model was then applied to the UAV images to produce two binary mask sets: one for vegetation and one for streets and sidewalks. These masks were incorporated during photogrammetric 3D reconstruction to estimate camera geometry and generate a dense point cloud. Our results show that vegetation masks did not improve camera geometry estimation. However, by applying UAV masks, we achieved a 15% reduction in total processing time and decreased the number of points by a factor of 2.7. This targeted approach enabled curb detection by focusing on expected curb locations. Curb candidate points were proposed using geometric characteristics of the point cloud, including normal values, linearity, and verticality. Our rule-based method effectively mapped even subtle curb features, providing a rapid, cost-effective solution for large-area curb mapping. Further, we explored the potential of random forest for curb mapping, with promising results. Our approach can support urban flood modeling efforts and strengthen urban resilience for flood-prone communities.

Photosynthetic responses to temperature across the tropics: a meta-analytic approach.

Tropical forests exchange more carbon dioxide (CO2) with the atmosphere than any other terrestrial biome. Yet, uncertainty in the projected carbon balance over the next century is roughly three-times greater for the tropics than other ecosystems. Our limited knowledge of tropical plant physiological responses, including photosynthetic, to climate change is a substantial source of uncertainty in our ability to forecast the global terrestrial carbon sink. We used a meta-analytic approach, focusing on tropical photosynthetic temperature responses, to address this knowledge gap. Our dataset, gleaned from 18 independent studies, included leaf-level light saturated photosynthetic (Asat) temperature responses from 108 woody species, with additional temperature parameters (35 species) and rates (250 species) of both maximum rates of electron transport (Jmax) and Rubisco carboxylation (Vcmax). We investigated how these parameters responded to mean annual temperature (MAT), temperature variability, aridity, and elevation, as well as also how responses differed among successional strategy, leaf habit, and light environment. Optimum temperatures for Asat (ToptA) and Jmax (ToptJ) increased with MAT but not for Vcmax (ToptV). Although photosynthetic rates were higher for "light" than "shaded" leaves, light conditions did not generate differences in temperature response parameters. ToptA did not differ with successional strategy, but early successional species had ~4 °C wider thermal niches than mid/late species. Semi-deciduous species had ~1 °C higher ToptA than broadleaf evergreen. Most global modeling efforts consider all tropical forests as a single "broadleaf evergreen" functional type, but our data show that tropical species with different leaf habits display distinct temperature responses that should be included in modeling efforts. This novel research will inform modeling efforts to quantify tropical ecosystem carbon cycling and provide more accurate representations of how these key ecosystems will respond to altered temperature patterns in the face of climate warming.

Aspects of Modeling Coal Enrichment Processes by Gravity Methods

This study examines the challenges associated with processing hard coal, with a specific focus on gravitational enrichment methods and the utilization of jigs for coal separation. The research involves the simulation and modeling of physical property distributions and the analysis of both the feed density distribution and the characteristics of the enrichment products. Findings indicate that the resultant density distributions are influenced not only by the gravitational enrichment process but also by the preceding procedures and the inherent properties of the coal, such as particle size, sulfur content, and ash content, all of which significantly affect the quality of the outcomes. In modeling and optimization efforts, the study emphasizes approximating grain density using selected statistical distributions—specifically, the Weibull, logistic, and Gaudin–Schuhmann–Andreyev (GSA) distributions—before and after the enrichment process. Statistical analyses demonstrate that the GSA distribution most accurately fits the grain density distribution in the feed, while the Weibull distribution provides the best approximation for the separation products. The quality of these approximations was assessed using the coefficient of determination (R2) and the Mean Squared Error (MSE). The best quality of approximation for feed was obtained by means of the GSA distribution function, and the MSE was approximately 3.1 for two analyzed values of feed flow rates. In the case of concentrates and tailings, the results are not unequivocal.

Real-time lava flow forecasting during the 2022 Mauna Loa eruption response

On November 27, 2022, Mauna Loa (Hawai‘i) erupted for the first time in ∼38 years, initially producing lava flows that covered the floor of its summit caldera, Moku‘āweoweo. Over the first 12 h following the summit eruption, four main fissures opened on Mauna Loa’s Northeast Rift Zone, with “fissure 3” quickly becoming the dominant source of lava flows. For the next 12 days, fissure 3 produced a 19-km-long lava flow to the north, crossing the Mauna Loa Weather Observatory access road and coming within 2.8 km of inundating the Daniel K. Inouye Highway (Saddle Road). Within 40 min of fissure 3 opening, inundation modeling efforts had begun. For the duration of the eruption, the computational fluid dynamics model Lava2d was run in real time, using flow front locations and other field observations to make sequential forecast improvements, eventually producing a set of models which accurately predicted the routing and arrival times of lava from fissure 3. These models were used to inform timing estimates of possible future inundation of the Saddle Road. As the eruption progressed, almost 4000 Lava2d models were made using high-performance computing resources, providing critical information on uncertainty in multi-week forecasts. To our knowledge, this was the first ever real-time physics-based ensemble lava flow modeling and forecasting effort. Here, we present a chronology of these real-time efforts, focusing on the successes and limitations of this approach.

Consumption Dynamics in Mixed-Income Neighborhoods with Connected Households

AbstractWe investigate the dynamics of household consumption in a setting in which households are connected across income classes. Low- and high-income households form preferences endogenously, conditional on their own and their neighbor’s past consumption. The modeling effort relies on a stochastic dynamic model of interdependent consumer choice in which the demand for commodities evolves according to a non-linear difference equation with stochastic initial states. The analysis targets a region of the parameter space that corresponds to salient features of a mixed-income neighborhood in which households are connected. Standard methods of asymptotic analysis of dynamic systems (e.g. bifurcation analysis) are combined with numerical simulation, statistical modelling of extreme events and statistical estimation techniques to investigate the dynamics. From the mathematical point of view, our analysis reveals the existence of intricate bifurcation pattern, coexistence of multiple attractors, complex basins and long transients. The essential economic finding states that key features of household consumption vary significantly in the influence the high-income households exert on the preference formation of the low-income households. In particular, we find that the prevalence of long transients, i.e. long waiting times before convergence to asymptotic states occur, is inversely related to the type of connectedness considered. We demonstrate that the dynamics of the consumption trajectory evolving over an extended time period before it settles on long-run simple consumption pattern, may not at all be captured by an asymptotic state. Thus, policies targeting the economies in mixed-income neighborhoods that are solely based on information about long-run consumption states, might trigger unwanted, unanticipated effects.

The distribution of the near-solar bound dust grains detected with Parker Solar Probe

Context. Parker Solar Probe (PSP) counts dust impacts in the near-solar region, but modeling effort is needed to understand the dust population’s properties. Aims. We aim to constrain the dust cloud’s properties based on the flux observed by PSP. Methods. We developed a forward model for the bound dust detection rates using the formalism of 6D phase space distribution of the dust. We applied the model to the location table of different PSP solar encounter groups. We explain some of the near-perihelion features observed in the data as well as the broader characteristic of the dust flux between 0.15 AU and 0.5 AU. We compare the measurements of PSP to the measurements of Solar Orbiter near 1 AU to expose the differences between the two spacecraft. Results. We found that the dust flux observed by PSP between 0.15 AU and 0.5 AU in post-perihelia can be explained by dust on bound orbits and is consistent with a broad range of orbital parameters, including dust on circular orbits. However, the dust number density as a function of the heliocentric distance and the scaling of detection efficiency with relative speed are important to explain the observed flux variation. The data suggest that the slope of differential mass distribution, δ, is between 0.14 and 0.49. The near-perihelion observations, however, show the flux maxima, which are inconsistent with the circular dust model, and additional effects may play a role. We found an indication that the sunward side of PSP is less sensitive to the dust impacts than PSP’s other surfaces. Conclusions. We show that the dust flux on PSP can be explained by noncircular bound dust and the detection capabilities of PSP. The scaling of flux with impact speed is especially important, and shallower than previously assumed.

Exploring deep learning models for roadside landslide prediction: Insights and implications from comparative analysis

This study undertakes a comparative analysis of four distinct deep learning models, i.e., Convolutional Neural Network (CNN), Deep Neural Network (DNN), Recurrent Neural Network (RNN), and Long Short-Term Memory (LSTM), in the context of roadside landslide prediction, aiming to provide comprehensive insights into their strengths and weaknesses. A geospatial dataset from the Lai Châu province, Vietnam, with thirteen environmental factors (elevation, aspect, slope, curvature, topographic wetness index, stream power index, flow accumulation, geology, normalized difference vegetation index, maximum rainfall, average annual rainfall, and proximity to faults and rivers) and 284 road-along landslides was considered for analysis. Our modeling efforts yielded invaluable insights into the performance of these models during both training and validation phases. The DNN model emerged as the frontrunner in the training phase, boasting the highest area under the curve (AUC) of 0.94, accuracy of 87.47%, kappa of 0.748, and lowest RMSE of 0.125. However, during validation, the CNN model outshone others, exhibiting the highest AUC of 0.88 and overall accuracy of 80.00%. Despite variations in performance metrics across phases, CNN consistently demonstrated robust predictive prowess. The findings of this study underscore the significance of selecting appropriate machine learning models tailored to specific contexts and objectives. Moreover, they contribute valuable insights for decision-makers and researchers alike, ultimately aiming to enhance the safety and resilience of communities inhabiting landslide-prone areas. Moving forward, future research directions may explore ensemble methods, novel architectures, and interpretability techniques to further advance predictive accuracy and applicability in roadside landslide susceptibility modeling.