Hasil untuk "Physical geography"

Giuseppe Vecchio, Adrien Kaiser, Rouffet Romain et al.

Foundation models have emerged as effective backbones for many vision tasks. However, current self-supervised features entangle high-level semantics with low-level physical factors, such as geometry and illumination, hindering their use in tasks requiring explicit physical reasoning. In this paper, we introduce $Φ$eat, a novel physically-grounded visual backbone that encourages a representation sensitive to material identity, including reflectance cues and geometric mesostructure. Our key idea is to employ a pretraining strategy that contrasts spatial crops and physical augmentations of the same material under varying shapes and lighting conditions. While similar data have been used in high-end supervised tasks such as intrinsic decomposition or material estimation, we demonstrate that a pure self-supervised training strategy, without explicit labels, already provides a strong prior for tasks requiring robust features invariant to external physical factors. We evaluate the learned representations through feature similarity analysis and material selection, showing that $Φ$eat captures physically-grounded structure beyond semantic grouping. These findings highlight the promise of unsupervised physical feature learning as a foundation for physics-aware perception in vision and graphics. These findings highlight the promise of unsupervised physical feature learning as a foundation for physics-aware perception in vision and graphics.

Congyue Deng, Brandon Y. Feng, Cecilia Garraffo et al.

Machine learning frameworks for physical problems must capture and enforce physical constraints that preserve the structure of dynamical systems. Many existing approaches achieve this by integrating physical operators into neural networks. While these methods offer theoretical guarantees, they face two key limitations: (i) they primarily model local relations between adjacent time steps, overlooking longer-range or higher-level physical interactions, and (ii) they focus on forward simulation while neglecting broader physical reasoning tasks. We propose the Denoising Hamiltonian Network (DHN), a novel framework that generalizes Hamiltonian mechanics operators into more flexible neural operators. DHN captures non-local temporal relationships and mitigates numerical integration errors through a denoising mechanism. DHN also supports multi-system modeling with a global conditioning mechanism. We demonstrate its effectiveness and flexibility across three diverse physical reasoning tasks with distinct inputs and outputs.

Hyunmin Cheong, Adrian Butscher

The current work presents an ontology developed for physics-based simulation in engineering design, called Physics-based Simulation Ontology (PSO). The purpose of the ontology is to assist in modelling the physical phenomenon of interest in a veridical manner, while capturing the necessary and reusable information for physics-based simulation solvers. The development involved extending an existing upper ontology, Basic Formal Ontology (BFO), to define lower-level terms of PSO. PSO has two parts: PSO-Physics, which consists of terms and relations used to model physical phenomena based on the perspective of classical mechanics involving partial differential equations, and PSO-Sim, which consists of terms used to represent the information artefacts that are about the physical phenomena modelled with PSO-Physics. The former terms are used to model the physical phenomenon of interest independent of solver-specific interpretations, which can be reused across different solvers, while the latter terms are used to instantiate solver-specific input data. A case study involving two simulation solvers was conducted to demonstrate this capability of PSO. Discussion around the benefits and limitations of using BFO for the current work is also provided, which should be valuable for any future work that extends an existing upper ontology to develop ontologies for engineering applications.

Ernesto Saiz, Faiz Iqbal, Jack H. Grant et al.

Field-deployable sensors play an important role in precision agriculture and are used to enable key stakeholders (farmers, agronomists, policy makers, etc.) to make informed decisions about resource management and improve crop yields. Sensor developers must consider technological, economic, and human aspects jointly in order to design a successful sensor. The objective of this review is to describe some of the key strands of each aspect and highlight the need to develop a degree of understanding of all of these aspects in order to create technologies that will be easily and readily adopted. Rather than analyzing each area in depth, we limit our discussion to a few major aspects and try to indicate interdependency and showcase their impact on one another. We hope that this approach will instigate thoughts and ideas and stimulate the co-creation of fit-for-purpose sensors.

Tao Liu, Ziqiang Zhang, Ping Du et al.

Effective identification of building patterns can improve the quality of automated map generalization. Established methods are limited by complex rule definitions and cannot fully consider the local domain information of buildings patterns, and few studies have focused on recognizing multiple building group patterns using GCN. Therefore, we propose a new model for recognizing building patterns by combining knowledge graph methods and GCN methods. First, the features of individual buildings are acquired, then the graph structure of building is constructed, and finally, by means of GCN and knowledge embedding, the features of buildings are efficiently learned on the basis of the graph structure of building patterns, and the pattern features of building are extracted, so as to realize the recognition of building patterns. The results show that the training accuracy and testing accuracy reach 99.03% and 95.89%, respectively. Compared with other methods, the proposed model can effectively utilize the local spatial information of building patterns and accurately recognize building patterns.

Wen Yin, Jian Lou, Pan Zhou et al.

Backdoor attacks have been well-studied in visible light object detection (VLOD) in recent years. However, VLOD can not effectively work in dark and temperature-sensitive scenarios. Instead, thermal infrared object detection (TIOD) is the most accessible and practical in such environments. In this paper, our team is the first to investigate the security vulnerabilities associated with TIOD in the context of backdoor attacks, spanning both the digital and physical realms. We introduce two novel types of backdoor attacks on TIOD, each offering unique capabilities: Object-affecting Attack and Range-affecting Attack. We conduct a comprehensive analysis of key factors influencing trigger design, which include temperature, size, material, and concealment. These factors, especially temperature, significantly impact the efficacy of backdoor attacks on TIOD. A thorough understanding of these factors will serve as a foundation for designing physical triggers and temperature controlling experiments. Our study includes extensive experiments conducted in both digital and physical environments. In the digital realm, we evaluate our approach using benchmark datasets for TIOD, achieving an Attack Success Rate (ASR) of up to 98.21%. In the physical realm, we test our approach in two real-world settings: a traffic intersection and a parking lot, using a thermal infrared camera. Here, we attain an ASR of up to 98.38%.

Yao Li, Pin Wang, Yihan Lou et al.

The increasing frequency of urban flood disasters presents a significant obstacle to urban sustainability. Urban flood management aims to reduce the flood occurrences, currently addressed through urban drainage systems. Previous studies have demonstrated future precipitation extremes will pose larger pressure on urban drainage network, but when and where the pressure will reach a dangerous level have never been assessed in any city of China. This study establishes the initial framework for identifying critical decades and hot spots of urban drainage pressure changes due to future climate change, through a case study conducted in southern China (Haining city). Urban drainage pressure was assessed by a combination of the urban drainage model known as the Storm Water Management Model (SWMM) and pipe statistics. Using climate projections from the latest phase of Coupled Model Intercomparison Project (CMIP6) under four typical SSP-RCP (shared socioeconomic pathway-representative concentration pathway) scenarios, we project the changes in urban drainage pressure by 21st century, and identify the key decades and high risk areas with the occurrence of dangerous pressure levels. The results indicate an overall upward trend in urban drainage pressure for Haining city, with over 97% of the flooding nodes projected to firstly reach the dangerous level by the 2030 s. Comparisons of the patterns under different SSP-RCP scenarios, suggest that a higher forcing pathway would expedite the deterioration of urban drainage pressure, particularly in urban areas with lower DEM and high building intensity. This has broad implications for better informing disaster management and policy-making in similar cities, especially those with inadequate drainage capacities.

Di Ma, Zhixu Bai, Yue-Ping Xu et al.

Study region: The Upper Mekong River Basin (UMRB), Southwest China. Study focus: With climate change unfolding and climate change knowledge evolving over time, it is imperative to investigate whether the latest CMIP6 models offer enhanced utility in climate change impact studies compared to their predecessors. This study strengthens the comparison between CMIP5 and CMIP6 models in assessing hydrological responses to future climate change. This was achieved utilizing the Soil and Water Assessment Tool, driven by downscaled CMIP5/CMIP6 model outputs under RCP8.5/SSP5–8.5. Both streamflow and sediment responses, encompassing the spatial and temporal changes, and the relationships between streamflow and sediment loads, were carefully evaluated and compared between CMIP5 and CMIP6. New hydrological insights for the region: CMIP6 indicates a stronger warming in 2071–2100 over the UMRB compared to CMIP5. Mean annual precipitation/streamflow is projected to increase by 22.7%/26.3% using CMIP5 and 28.4%/34.4% using CMIP6. Mean annual sediment load changes, however, show a discrepancy between CMIP5 (−3.7%) and CMIP6 (+13.8%). CMIP6 exhibits larger inter-model variability in both climate and hydrological projections. Regarding future spatial distributions of mean annual water and sediment yields, a considerable agreement is demonstrated between CMIP5 and CMIP6, despite CMIP6 showing larger projections over most subbasins. Additionally, both ensembles exhibit approximate relationships between streamflow and sediment loads, indicating a comparable decline in watershed sediment generation and transport capacity under future climate change. Overall, CMIP6 suggests more severe climate change impacts on streamflow and sediment loads in the UMRB than CMIP5, highlighting the need to update climate change adaptation and mitigation policies based on the latest insights derived from CMIP6.

Floyd W. Stecker

This book chapter presents an overview of the historical experimental and theoretical developments in neutrino physics and astrophysics and also the physical properties of neutrinos, as well as the physical processes involving neutrinos. It also discusses the role of neutrinos in astrophysics and cosmology. Correction to tex file made.

Feng Gao, Jyoti Jennewein, W. Dean Hively et al.

Cover crops are planted to reduce soil erosion, increase soil fertility, and improve watershed management. In the Delmarva Peninsula of the eastern United States, winter cover crops are essential for reducing nutrient and sediment losses from farmland. Cost-share programs have been created to incentivize cover crops to achieve conservation objectives. This program required that cover crops be planted and terminated within a specified time window. Usually, farmers report cover crop termination dates for each enrolled field (∼28,000 per year), and conservation district staff confirm the report with field visits within two weeks of termination. This verification process is labor-intensive and time-consuming and became restricted in 2020–2021 due to the COVID-19 pandemic. This study used Harmonized Landsat and Sentinel-2 (HLS, version 2.0) time-series data and the within-season termination (WIST) algorithm to detect cover crop termination dates over Maryland and the Delmarva Peninsula. The estimated remote sensing termination dates were compared to roadside surveys and to farmer-reported termination dates from the Maryland Department of Agriculture database for the 2020–2021 cover crop season. The results show that the WIST algorithm using HLS detected 94% of terminations (statuses) for the enrolled fields (n = 28,190). Among the detected terminations, about 49%, 72%, 84%, and 90% of remote sensing detected termination dates were within one, two, three, and four weeks of agreement to farmer-reported dates, respectively. A real-time simulation showed that the termination dates could be detected one week after termination operation using routinely available HLS data, and termination dates detected after mid-May are more reliable than those from early spring when the Normalized Difference Vegetation Index (NDVI) was low. We conclude that HLS imagery and the WIST algorithm provide a fast and consistent approach for generating near-real-time cover crop termination maps over large areas, which can be used to support cost-share program verification.

Shiqian Li, Kewen Wu, Chi Zhang et al.

Is dynamics prediction indispensable for physical reasoning? If so, what kind of roles do the dynamics prediction modules play during the physical reasoning process? Most studies focus on designing dynamics prediction networks and treating physical reasoning as a downstream task without investigating the questions above, taking for granted that the designed dynamics prediction would undoubtedly help the reasoning process. In this work, we take a closer look at this assumption, exploring this fundamental hypothesis by comparing two learning mechanisms: Learning from Dynamics (LfD) and Learning from Intuition (LfI). In the first experiment, we directly examine and compare these two mechanisms. Results show a surprising finding: Simple LfI is better than or on par with state-of-the-art LfD. This observation leads to the second experiment with Ground-truth Dynamics, the ideal case of LfD wherein dynamics are obtained directly from a simulator. Results show that dynamics, if directly given instead of approximated, would achieve much higher performance than LfI alone on physical reasoning; this essentially serves as the performance upper bound. Yet practically, LfD mechanism can only predict Approximate Dynamics using dynamics learning modules that mimic the physical laws, making the following downstream physical reasoning modules degenerate into the LfI paradigm; see the third experiment. We note that this issue is hard to mitigate, as dynamics prediction errors inevitably accumulate in the long horizon. Finally, in the fourth experiment, we note that LfI, the extremely simpler strategy when done right, is more effective in learning to solve physical reasoning problems. Taken together, the results on the challenging benchmark of PHYRE show that LfI is, if not better, as good as LfD for dynamics prediction. However, the potential improvement from LfD, though challenging, remains lucrative.

Shouzhi Chen, Yongshuo H. Fu, Fanghua Hao et al.

The Earth is experiencing unprecedented climate change. Vegetation phenology has already showed strong response to the global warming, which alters mass and energy fluxes on terrestrial ecosystems. With technology and method developments in remote sensing, computer science and citizen science, many recent phenology-related studies have been focused on macrophenology. In this perspective, we 1) reviewed the responses of vegetation phenology to climate change and its impacts on carbon cycling, and reported that the effect of shifted phenology on the terrestrial carbon fluxes is substantially different between spring and autumn; 2) elaborated how vegetation phenology affects ecohydrological processes at different scales, and further listed the key issues for each scale, i.e., focusing on seasonal effect, local feedbacks and regional vapor transport for individual, watershed and global respectively); 3) envisioned the potentials to improve current hydrological models by coupling vegetation phenology-related processes, in combining with machine learning, deep learning and scale transformation methods. We propose that comprehensive understanding of climate-macrophenology-hydrology interactions are essential and urgently needed for enhancing our understanding of the ecosystem response and its role in hydrological cycle under future climate change.

Cheng Xue, Vimukthini Pinto, Chathura Gamage et al.

Reasoning about the behaviour of physical objects is a key capability of agents operating in physical worlds. Humans are very experienced in physical reasoning while it remains a major challenge for AI. To facilitate research addressing this problem, several benchmarks have been proposed recently. However, these benchmarks do not enable us to measure an agent's granular physical reasoning capabilities when solving a complex reasoning task. In this paper, we propose a new benchmark for physical reasoning that allows us to test individual physical reasoning capabilities. Inspired by how humans acquire these capabilities, we propose a general hierarchy of physical reasoning capabilities with increasing complexity. Our benchmark tests capabilities according to this hierarchy through generated physical reasoning tasks in the video game Angry Birds. This benchmark enables us to conduct a comprehensive agent evaluation by measuring the agent's granular physical reasoning capabilities. We conduct an evaluation with human players, learning agents, and heuristic agents and determine their capabilities. Our evaluation shows that learning agents, with good local generalization ability, still struggle to learn the underlying physical reasoning capabilities and perform worse than current state-of-the-art heuristic agents and humans. We believe that this benchmark will encourage researchers to develop intelligent agents with advanced, human-like physical reasoning capabilities. URL: https://github.com/Cheng-Xue/Hi-Phy

Kaushal K. Garg, K.H. Anantha, Rajesh Nune et al.

Study region: This study analyzes the impact of land use changes on the hydrology of Kolar district in the state of Karnataka, India. Kolar receives on average 565 mm (σ = 130) rainfall during June to October and has a wide gap between its water supply and demand. Study focus: This research identifies the reasons and causes of the gap. A water balance model was successfully calibrated and validated against measurements of groundwater level, recharge and surface runoff. New hydrological insights for the region: The study revealed that between 1972 and 2011, there was a major shift from grass and rainfed crop lands to eucalyptus plantation and irrigated cultivation. About 17.7 % and 18 % of the district area converted into eucalyptus plantation and irrigated lands during this period, respectively. Eucalyptus plantations tended to cause large losses by ET leading to increase in soil moisture deficit and reduction in the recharge to groundwater and in surface runoff (approx. 30 %). The irrigation demand of the district increased from 57 mm (1972) to 140 mm (2011) which resulted in increased groundwater abstraction by 145 %. The expansion of the irrigated area is the major contributing factor for widening the demand-supply gap (62 %) of the freshwater availability. Results could help various stakeholders, including district and national authorities to develop the most suitable water management strategies in order to close the gap between water supply and demand.

Craig Lage

Removing systematic effects from astronomical images taken with CCDs requires a detailed understanding of the physics of the imaging process. To aid in this understanding, we have built detailed electrostatic simulations of the LSST CCDs. In order to build an electrostatic model of the LSST CCDs, physical information about the CCDs is required. These details include things such as the physical dimensions of the components of the CCD, dopant profiles, and in some cases, electrical measurements of the CCD. This work documents the results of these physical and electrical measurements on LSST CCDs.

Maxim Yu. Khlopov

The lack of experimental evidence at the LHC for physics beyond the Standard model (BSM) of elementary particles together with necessity of its existence to provide solutions of internal problems of the Standard model (SM) as well as of physical nature of the basic elements of the modern cosmology demonstrates the conspiracy of BSM physics. Simultaneously the data of precision cosmology only tighten the constraints on the deviations from the now standard LambdaCDM model and thus exhibit conspiracy of the nonstandard cosmological scenarios. We show that studying new physics in combination of its physical, astrophysical and cosmological probes, can not only unveil the conspiracy of BSM physics but will also inevitably reveal nonstandard features in the cosmological scenario.

Robert Geroch, James B. Hartle

The familiar theories of physics have the feature that the application of the theory to make predictions in specific circumstances can be done by means of an algorithm. We propose a more precise formulation of this feature --- one based on the issue of whether or not the physically measurable numbers predicted by the theory are computable in the mathematical sense. Applying this formulation to one approach to a quantum theory of gravity, there are found indications that there may exist no such algorithms in this case. Finally, we discuss the issue of whether the existence of an algorithm to implement a theory should be adopted as a criterion for acceptable physical theories.

Kata Szilágyi

This article summarizes the current state of research on the flaked stone assemblages from the Late Neolithic site Alsónyék‒Bátaszék, Tolna district. The raw material distribution of the nearly 6100 pieces that make up the stone tool assemblage is the focus of this paper, with a particular emphasis placed on the dominance of the local raw material. The research addresses the question of the method of procurement of the lithic raw material in the case of this enormous, extended Neolithic site. To supply an answer, basic geoarchaeological research was necessary. To that end, a field survey aimed at detecting those geological formations and lithic variations convenient for knapping was undertaken. The results of the survey reported in the second part of this paper help in our understanding of the selection strategy of the ancient knapping specialists. From these strategies, it is possible to recognize the cultural tradition and raw material manipulation of this Late Neolithic community and, in a wider sense, the southeastern group of the Lengyel culture

Ori Adam

Abstract In the present climate, tropical rain bands exhibit a bifurcated pattern, continuously forming along single intertropical convergence zones (ITCZs) in some regions, and along double ITCZs that straddle the equator in other regions. This bifurcated ITCZ pattern is studied in a Matsuno‐Gill‐type model forced by relaxation to zonally asymmetric surface pressure. The model includes an idealized Bjerknes feedback which couples surface winds and sea surface temperatures (SSTs) via oceanic Ekman balance. Consistent with observations, solutions in the limit of strong damping are explored. Two ITCZ bifurcation mechanisms are identified. First, in the viscous limit, ITCZs form along negative anomalies of the local Rossby number, which tend to occur near the equator for equatorial low pressure and off the equator for equatorial high pressure; this leads to a single ITCZ in the rising branch of zonal‐overturning circulations and a double ITCZ that straddles the equator in the descending branch. Second, ocean upwelling produces an equatorial cold tongue with increased surface pressure, which reduces vertical winds and can lead to precipitation peaks that straddle the equator in regions of near‐equator ascent. Consistent with observations, the cold tongue intensifies with increasing zonal SST gradients, and its base widens with weakened poleward SST gradients, modulating the zonal orientation of the ITCZs on either side of the cold tongue. Analytic approximate solutions in the viscous limit capture the emergence of the bifurcated ITCZ pattern, as well as the dependence of the bifurcated ITCZ pattern on zonal and poleward SST gradients

Sebastien Ehrhardt, Aron Monszpart, Niloy J. Mitra et al.

Evolution has resulted in highly developed abilities in many natural intelligences to quickly and accurately predict mechanical phenomena. Humans have successfully developed laws of physics to abstract and model such mechanical phenomena. In the context of artificial intelligence, a recent line of work has focused on estimating physical parameters based on sensory data and use them in physical simulators to make long-term predictions. In contrast, we investigate the effectiveness of a single neural network for end-to-end long-term prediction of mechanical phenomena. Based on extensive evaluation, we demonstrate that such networks can outperform alternate approaches having even access to ground-truth physical simulators, especially when some physical parameters are unobserved or not known a-priori. Further, our network outputs a distribution of outcomes to capture the inherent uncertainty in the data. Our approach demonstrates for the first time the possibility of making actionable long-term predictions from sensor data without requiring to explicitly model the underlying physical laws.