Article
Computer Science, Artificial Intelligence
Bi Wang, Jianqing Wu, Xuelian Li, Jun Shen, Yangjun Zhong
Summary: In this paper, a variant of Q-learning, named uncertainty quantification based Q-learning, is proposed by introducing the hedonistic expected value (HEV) to increase the probability of outputting an optimal partial order in online reinforcement learning. The weights assigned by HEV to the successors are compatible with the existing operators, and the prediction of the return is not only the sum over the weights succeeding the operator but also over the weights following HEV through re-weighting. The proposed algorithm with HEV demonstrates favorable performance in practice.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Geosciences, Multidisciplinary
S. Peatier, B. M. Sanderson, L. Terray, R. Roehrig
Summary: This study samples 30 calibration parameters of the CNRM-CM6-1 atmospheric component to create a perturbed parameter ensemble. By using statistical emulators of climatic fields and feedback parameters, the study proposes a method to generate model candidates that cover the maximal range of net feedback while minimizing the error score. The optimal candidates have large errors in the extremes of climate sensitivity but fall within the inter-model standard deviation of other models participating in the Cloud Feedback Model Intercomparison Project.
GEOPHYSICAL RESEARCH LETTERS
(2022)
Article
Ecology
Jody R. Reimer, Frederick R. Adler, Kenneth M. Golden, Akil Narayan
Summary: Uncertainty in parameters in ecological models can be incorporated by treating parameters as random variables with distributions. Recent advances in uncertainty quantification methods provide new approaches for analyzing models with random parameters. Modelling key parameters as random variables changes the characteristics of the model. The computational efficiency of polynomial chaos methods helps in better predicting and synthesizing models with data.
Article
Engineering, Electrical & Electronic
Niloofar Rashidi, Qiong Wang, Rolando Burgos, Chris Roy, Dushan Boroyevich
Summary: This article introduces a robust multi-objective design and optimization approach with parametric and model-form uncertainty quantification, highlighting the benefits of incorporating parametric uncertainty quantification into the design optimization framework to reduce design sensitivity and improve the effectiveness of optimal design solutions.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2021)
Article
Engineering, Aerospace
Ziming Song, Zaijie Liu, Jiachen Lu, Chao Yan
Summary: Due to the lack of physical knowledge of boundary layer transition, the y-Reo transition model introduces closure parameters, which increase the uncertainty of transition prediction. This work quantifies the uncertainties of closure parameters in the quantities of interests and identifies the key parameters. The relative contribution of each parameter to uncertainty is evaluated by the Sobol index. The results show that ce2 and ca2 are the key parameters of the y-Reo model.
CHINESE JOURNAL OF AERONAUTICS
(2023)
Article
Mathematics, Applied
Moritz Kassmann, Kim Kyung-Youn, Takashi Kumagai
Summary: We prove sharp two-sided bounds of the fundamental solution for integro-differential operators of order alpha is an element of (0, 2) that generate a d-dimensional Markov process. The corresponding Dirichlet form is comparable to that of d independent copies of one-dimensional jump processes, i.e., the jumping measure is singular with respect to the d-dimensional Lebesgue measure.
JOURNAL DE MATHEMATIQUES PURES ET APPLIQUEES
(2022)
Article
Chemistry, Physical
Yonatan Kurniawan, Cody L. Petrie, Kinamo J. J. Williams, Mark K. Transtrum, Ellad B. Tadmor, Ryan S. Elliott, Daniel S. Karls, Mingjian Wen
Summary: This paper investigates the quantification of parametric uncertainty in classical empirical interatomic potentials using Bayesian and frequentist methods. It reveals that these potentials are typically insensitive and parameters are unidentifiable. Information geometry is used to explain the underlying cause and suggest new parameterizations and simplified models.
JOURNAL OF CHEMICAL PHYSICS
(2022)
Article
Engineering, Mechanical
Marco Behrendt, Matthias G. R. Faes, Marcos A. Valdebenito, Michael Beer
Summary: In engineering, the modelling of environmental processes is essential for designing structures safely and determining the reliability of existing structures. This work focuses on situations where data is limited and it is not feasible to derive reliable statistics. The proposed approach uses a radial basis function network to generate basis functions that enclose the data, resulting in an interval-based power spectral density (PSD) function. The applicability of this imprecise PSD model is demonstrated with recorded earthquake ground motions.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Physics, Multidisciplinary
Jesni Shamsul Shaari, Rinie N. M. Nasir, Stefano Mancini
Summary: We reformulated the notion of uncertainty for pairs of unitary operators in the context of guessing games and derived an entropic uncertainty relation for such pairs. We demonstrated the compatibility of distinguishable operators and connected maximal incompatibility of unitary operators to bases for some subspace of mutually unbiased operators.
Article
Green & Sustainable Science & Technology
Ehsan Ranaee, Rafi Khattar, Fabio Inzoli, Martin J. Blunt, Alberto Guadagnini
Summary: We conducted a probabilistic assessment of CO2 storage capacity in major sedimentary basins in China, showing that 10 major basins have the potential to store an average of 1350 Gt of CO2 in the next 30 years. Our analysis suggests that underground carbon storage in China, combined with other solutions, could help meet the goals of mitigating global warming by 2060, as per the Announced Pledges Scenario (International Energy Agency). Geological formation attributes were found to be major sources of uncertainty affecting model outputs and associated uncertainty.
INTERNATIONAL JOURNAL OF GREENHOUSE GAS CONTROL
(2022)
Article
Engineering, Mechanical
Benjamin Froehlich, Dominik Hose, Oliver Dieterich, Michael Hanss, Peter Eberhard
Summary: The finite element method is a widely used tool for modeling complex engineering structures. However, the uncertainty in model input parameters can significantly affect simulation results. This article proposes a novel simulation workflow that combines parametric modeling and parametric model order reduction to efficiently quantify uncertainties in large-scale dynamical systems.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Automation & Control Systems
Kai Shen, Dillard L. Robertson, Joseph K. Scott
Summary: This paper presents a new method for tighter enclosures of trajectories consistent with a system of nonlinear ordinary differential equations by extending the DI approach with mean value form. Mean-value DI method converges quadratically to zero overestimation error as uncertainty set diminishes, which is crucial for algorithms based on uncertainty set partitioning. Additionally, numerical results for challenging test problems are provided to demonstrate the effectiveness of the proposed method.
Article
Multidisciplinary Sciences
Charlotte J. Alster, Allycia van de Laar, Jordan P. Goodrich, Vickery L. Arcus, Julie R. Deslippe, Alexis J. Marshall, Louis A. Schipper
Summary: Quantifying the rate of thermal adaptation of soil microbial respiration is crucial for understanding the potential impact of carbon cycle feedbacks under a warming climate. This study found that microbial respiration exhibited an adaptation rate related to warming.
NATURE COMMUNICATIONS
(2023)
Article
Environmental Sciences
Tina Karimi, Patrick Reed, Keyvan Malek, Jennifer Adam
Summary: This study clarifies how parametric uncertainty in agro-hydrologic models influences yield projections under changing future climate. It also highlights the potential bias introduced by using stationarity assumptions in calibrating model parameters and making future yield projections.
WATER RESOURCES RESEARCH
(2022)
Article
Biophysics
Jeremy A. Owen, Pranay Talla, John W. Biddle, Jeremy Gunawardena
Summary: Switch-like motifs, particularly ultrasensitive switches, which consist of two enzymes acting antagonistically on a substrate by making or removing a covalent modification, play important roles in biochemical networks. In this study, the linear framework for timescale separation was used to establish strict bounds on the performance of any covalent-modification switch in terms of the chemical potential difference driving the cycle. These bounds apply to different enzyme mechanisms and rate constants, providing fundamental physical constraints on covalent switching.
BIOPHYSICAL JOURNAL
(2023)
Article
Geosciences, Multidisciplinary
Tom Dror, Vered Silverman, Orit Altaratz, Mickael D. Chekroun, Ilan Koren
Summary: This article investigates the limitations in representing shallow cumulus in climate prediction, and identifies a class of continental shallow convective cumulus clouds that share distinct morphological properties. These clouds, named greenCu, are mainly found over forests and vegetated areas. Through the creation of a satellite-based dataset, the study reveals that greenCu clouds are driven by similar large-scale meteorological conditions regardless of their geographical locations across continents.
GEOPHYSICAL RESEARCH LETTERS
(2022)
Correction
Multidisciplinary Sciences
Stefano Pierini, Michael Ghil
SCIENTIFIC REPORTS
(2022)
Article
Geosciences, Multidisciplinary
Damian Walwer, Michael Ghil, Eric Calais
Summary: By studying space geodetic time series and using M-SSA methodology, it is possible to extract qualitative dynamics information from deformation data of volcanic systems, aiding in the understanding of pressure build-up within magma bodies and providing guidelines for physics-based models of episodic inflation cycles.
FRONTIERS IN EARTH SCIENCE
(2022)
Article
Meteorology & Atmospheric Sciences
J. David Neelin, Cristian Martinez-Villalobos, Samuel N. Stechmann, Fiaz Ahmed, Gang Chen, Jesse M. Norris, Yi-Hung Kuo, Geert Lenderink
Summary: This review examines the relationship between precipitation and its thermodynamic environment, specifically water vapor and temperature, and the implications for extreme precipitation changes in a warmer climate. It discusses empirical relationships, changes in precipitation extremes under warming scenarios, and fundamental processes that influence precipitation distributions. The findings suggest that while water vapor increases are governed by temperature, precipitation extreme changes are more complex and can increase rapidly, particularly in the tropics. Integrating different research threads and developing a comprehensive explanation of precipitation probability distribution could advance the understanding in this field.
CURRENT CLIMATE CHANGE REPORTS
(2022)
Article
Meteorology & Atmospheric Sciences
Todd Emmenegger, Yi-Hung Kuo, Shaocheng Xie, Chengzhu Zhang, Cheng Tao, J. David Neelin
Summary: A set of diagnostics is used to assess the behavior of the Coupled Model Intercomparison Project (CMIP6) models with respect to precipitation. The models show significant errors in the relationship between precipitation and column water vapor (CWV). Models also exhibit biases in column relative humidity (CRH) statistics, with compensating biases often occurring.
JOURNAL OF CLIMATE
(2022)
Article
Meteorology & Atmospheric Sciences
Yongkang Xue, Ismaila Diallo, Aaron A. Boone, Tandong Yao, Yang Zhang, Xubin Zeng, J. David Neelin, William K. M. Lau, Yan Pan, Ye Liu, Xiaoduo Pan, Qi Tang, Peter J. van Oevelen, Tomonori Sato, Myung-Seo Koo, Stefano Materia, Chunxiang Shi, Jing Yang, Constantin Ardilouze, Zhaohui Lin, Xin Qi, Tetsu Nakamura, Subodh K. Saha, Retish Senan, Yuhei Takaya, Hailan Wang, Hongliang Zhang, Mei Zhao, Hara Prasad Nayak, Qiuyu Chen, Jinming Feng, Michael A. Brunke, Tianyi Fan, Songyou Hong, Paulo Nobre, Daniele Peano, Yi Qin, Frederic Vitart, Shaocheng Xie, Yanling Zhan, Daniel Klocke, Ruby Leung, Xin Li, Michael Ek, Weidong Guo, Gianpaolo Balsamo, Qing Bao, Sin Chan Chou, Patricia de Rosnay, Yanluan Lin, Yuejian Zhu, Yun Qian, Ping Zhao, Jianping Tang, Xin-Zhong Liang, Jinkyu Hong, Duoying Ji, Zhenming Ji, Yuan Qiu, Shiori Sugimoto, Weicai Wang, Kun Yang, Miao Yu
Summary: The surface temperature of the Tibetan Plateau is causally related to summer precipitation in multiple regions across the world, indicating that high-mountain land temperature could be a substantial source of subseasonal-to-seasonal precipitation predictability.
BULLETIN OF THE AMERICAN METEOROLOGICAL SOCIETY
(2022)
Article
Astronomy & Astrophysics
Dmitri Kondrashov, Alexander Y. Drozdov, Daniel Vech, David M. Malaspina
Summary: We propose a random forests machine learning model for accurate prediction of plasmaspheric hiss spectral classes. The model's high predictive skill is largely determined by the spatial location of each spectral class, while the inclusion of other predictors only provides minor improvements in predictive skill.
FRONTIERS IN ASTRONOMY AND SPACE SCIENCES
(2022)
Article
Mathematics
Mickael D. Chekroun, Honghu Liu, James C. McWilliams, Shouhong Wang
Summary: A central challenge in physics is to describe non-equilibrium systems driven by randomness. In this study, an alternative framework is presented to cope with the difficulties of characterizing systems with random fluctuations. The approach employs the approximation theory of stochastic invariant manifolds and energy estimates to predict the occurrence of stochastic bifurcation in reduced equations of stochastic fluid problems. Applications to a stochastic Rayleigh-Benard problem are detailed, clarifying conditions for a stochastic pitchfork bifurcation to occur.
JOURNAL OF DIFFERENTIAL EQUATIONS
(2023)
Article
Mathematics, Applied
Dan Crisan, Michael Ghil
Summary: Extensive numerical evidence demonstrates that assimilating observations has a stabilizing effect on unstable dynamics, both in numerical weather prediction and other fields. In this paper, we employ mathematically rigorous methods to explain the underlying reasons behind this phenomenon. Our stabilization results do not necessitate a complete set of observations and we provide examples where observing only the unstable degrees of freedom of the model is sufficient.
Article
Multidisciplinary Sciences
Witold Bagniewski, Denis-Didier Rousseau, Michael Ghil
Summary: Tipping points in Earth's climate system have become increasingly important due to the potential risk of abrupt, irreversible climate transitions caused by human activities. Paleoclimate records are crucial for identifying past tipping points and understanding the underlying mechanisms. However, the quality, resolution, and reliability of these records vary, making careful selection necessary. To address this, the open-source PaleoJump database offers high-resolution records from various sources and statistical methodologies for identifying and analyzing tipping points. This database provides a valuable resource for researchers studying past climate changes.
SCIENTIFIC REPORTS
(2023)
Article
Astronomy & Astrophysics
Alexander Y. Drozdov, Dmitri Kondrashov, Kirill Strounine, Yuri Y. Shprits
Summary: In this study, we reconstruct radiation belt electron fluxes using data assimilation with Polar Orbiting Environmental Satellites (POES) measurements mapped to near equatorial regions. We compare two machine learning methods (multivariate linear regression and neural network) for mapping POES measurements. The results show that the MLR-based mapping method provides a reasonably good agreement with observations, while the NN-based method performs better. However, the improvement by adding data assimilation is limited compared to the purely NN model.
FRONTIERS IN ASTRONOMY AND SPACE SCIENCES
(2023)
Article
Mathematics, Applied
Gisela D. Charo, Michael Ghil, Denisse Sciamarella
Summary: Random attractors are time-evolving structures of chaotic and perturbed dynamical systems that can be described using templex. Templex consists of cell complexes and directed graphs, providing a detailed description and classification of the attractors. Critical points in random templex lead to drastic changes in attractor holes over time.
Article
Environmental Sciences
Huan Liu, Ilan Koren, Orit Altaratz, Mickael D. Chekroun
Summary: Clouds have a crucial impact on Earth's energy balance and water cycle. Their response to global warming is the biggest source of uncertainty in climate prediction. By analyzing 42 years of global cloud coverage reanalysis data, we have identified a clear trend and El Nino-Southern Oscillation-related patterns. The trend reveals decreasing cloud coverage over most continents and increasing coverage over tropical and subtropical oceans. The reduction in near-surface relative humidity can explain the declining cloud coverage over land. Our findings suggest potential impacts on the terrestrial water cycle and changes in energy partitioning between land and ocean, all linked to global warming.
ATMOSPHERIC CHEMISTRY AND PHYSICS
(2023)
Article
Geosciences, Multidisciplinary
Keno Riechers, Takahito Mitsui, Niklas Boers, Michael Ghil
Summary: The relative importance of external forcing and intrinsic variability in climate change is a key question in understanding both general climate variability and the paleoclimatic history of our planet. Research over the past century has established the significance of orbital forcing in the last 2.6 million years and the Quaternary glaciation cycles. Internal mechanisms have also been found to play a causal role in events such as Dansgaard-Oeschger and Heinrich events as well as the mid-Pleistocene transition. This study introduces a unified framework that utilizes the theory of non-autonomous and random dynamical systems to understand the effects of orbital forcing on the climate system's internal variability over timescales ranging from thousands to millions of years.
CLIMATE OF THE PAST
(2022)
Article
Geosciences, Multidisciplinary
Denis-Didier Rousseau, Witold Bagniewski, Michael Ghil
Summary: This paper re-examines the climate variations determined from marine and Greenland records, and finds that these abrupt climate changes are related to the growth and retreat of ice sheets and the astronomical theory of climate.
CLIMATE OF THE PAST
(2022)
