Observational Uncertainties Research Articles

Exploring the limits of the Jenkinson–Collison weather types classification scheme: a global assessment based on various reanalyses

The Jenkinson–Collison weather typing scheme (JC-WT) is an automated method used to classify regional sea-level pressure into a reduced number of typical recurrent patterns. Originally developed for the British Isles in the early 1970’s on the basis of expert knowledge, the method since then has seen many applications. Encouraged by the premise that the JC-WT approach can in principle be applied to any mid-to-high latitude region, the present study explores its global extra-tropical applicability, including the Southern Hemisphere. To this aim, JC-WT is applied at each grid-box of a global 2.5^{circ } regular grid excluding the inner tropics (± 5^{circ } band). Thereby, 6-hourly JC-WT catalogues are obtained for 5 distinct reanalyses, covering the period 1979–2005, which are then applied to explore (1) the limits of method applicability and (2) observational uncertainties inherent to the reanalysis datasets. Using evaluation criteria, such as the diversity of occurring circulation types and the frequency of unclassified situations, we extract empirically derived applicability thresholds which suggest that JC-WT can be generally used anywhere polewards of 23.5^{circ }, with some exceptions. Seasonal fluctuations compromise this finding along the equatorward limits of the domain. Furthermore, unreliable reanalysis sea-level pressure estimates in elevated areas with complex orography (such as the Tibetan Plateau, the Andes, Greenland and Antarctica) prevent the application of the method in these regions. In some other regions, the JC-WT classifications obtained from the distinct reanalyses substantially differ from each other, which may bring additional uncertainties when the method is used in model evaluation experiments.

Cloud Climatologies from Global Climate Models—A Comparison of CMIP5 and CMIP6 Models with Satellite Data

Abstract Simulating clouds with global climate models is challenging as the relevant physics involves many nonlinear processes covering a wide range of spatial and temporal scales. As key components of the hydrological cycle and the climate system, an evaluation of clouds from models used for climate projections is an important prerequisite for assessing the confidence in the results from these models. Here, we compare output from models contributing to phase 6 of the Coupled Model Intercomparison Project (CMIP6) with satellite data and with results from their predecessors (CMIP5). We use multiproduct reference datasets to estimate the observational uncertainties associated with different sensors and with internal variability on a per-pixel basis. Selected cloud properties are also analyzed by region and by dynamical regime and thermodynamic conditions. Our results show that for parameters such as total cloud cover, cloud water path, and cloud radiative effect, the CMIP6 multimodel mean performs slightly better than the CMIP5 ensemble mean in terms of mean bias, pattern correlation, and relative root-mean square deviation. The intermodel spread in CMIP6, however, is not reduced compared to CMIP5. Compared with CALIPSO-ICECLOUD data, the CMIP5/6 models overestimate cloud ice, particularly in the lower and middle troposphere, partly due to too high ice fractions for given temperatures. This bias is reduced in the CMIP6 multimodel mean. While many known biases such as an underestimation in cloud cover in stratocumulus regions remain in CMIP6, we find that the CMIP5 problem of too few but too reflective clouds over the Southern Ocean is significantly improved.

Daedalus MASE (mission assessment through simulation exercise): A toolset for analysis of in situ missions and for processing global circulation model outputs in the lower thermosphere-ionosphere

Daedalus MASE (Mission Assessment through Simulation Exercise) is an open-source package of scientific analysis tools aimed at research in the Lower Thermosphere-Ionosphere (LTI). It was created with the purpose to assess the performance and demonstrate closure of the mission objectives of Daedalus, a mission concept targeting to perform in-situ measurements in the LTI. However, through its successful usage as a mission-simulator toolset, Daedalus MASE has evolved to encompass numerous capabilities related to LTI science and modeling. Inputs are geophysical observables in the LTI, which can be obtained either through in-situ measurements from spacecraft and rockets, or through Global Circulation Models (GCM). These include ion, neutral and electron densities, ion and neutral composition, ion, electron and neutral temperatures, ion drifts, neutral winds, electric field, and magnetic field. In the examples presented, these geophysical observables are obtained through NCAR’s Thermosphere-Ionosphere-Electrodynamics General Circulation Model. Capabilities of Daedalus MASE include: 1) Calculations of products that are derived from the above geophysical observables, such as Joule heating, energy transfer rates between species, electrical currents, electrical conductivity, ion-neutral collision frequencies between all combinations of species, as well as height-integrations of derived products. 2) Calculation and cross-comparison of collision frequencies and estimates of the effect of using different models of collision frequencies into derived products. 3) Calculation of the uncertainties of derived products based on the uncertainties of the geophysical observables, due to instrument errors or to uncertainties in measurement techniques. 4) Routines for the along-orbit interpolation within gridded datasets of GCMs. 5) Routines for the calculation of the global coverage of an in situ mission in regions of interest and for various conditions of solar and geomagnetic activity. 6) Calculations of the statistical significance of obtaining the primary and derived products throughout an in situ mission’s lifetime. 7) Routines for the visualization of 3D datasets of GCMs and of measurements along orbit. Daedalus MASE code is accompanied by a set of Jupyter Notebooks, incorporating all required theory, references, codes and plotting in a user-friendly environment. Daedalus MASE is developed and maintained at the Department for Electrical and Computer Engineering of the Democritus University of Thrace, with key contributions from several partner institutions.

Noise-Net: determining physical properties of H iiregions reflecting observational uncertainties

ABSTRACTStellar feedback, the energetic interaction between young stars and their birthplace, plays an important role in the star formation history of the Universe and the evolution of the interstellar medium. Correctly interpreting the observations of star-forming regions is essential to understand stellar feedback, but it is a non-trivial task due to the complexity of the feedback processes and degeneracy in observations. In our recent paper, we introduced a conditional invertible neural network (cINN) that predicts seven physical properties of star-forming regions from the luminosity of 12 optical emission lines as a novel method to analyse degenerate observations. We demonstrated that our network, trained on synthetic star-forming region models produced by the warpfield-emission predictor (warpfield-emp), could predict physical properties accurately and precisely. In this paper, we present a new updated version of the cINN that takes into account the observational uncertainties during network training. Our new network named Noise-Net reflects the influence of the uncertainty on the parameter prediction by using both emission-line luminosity and corresponding uncertainties as the necessary input information of the network. We examine the performance of the Noise-Net as a function of the uncertainty and compare it with the previous version of the cINN, which does not learn uncertainties during the training. We confirm that the Noise-Net outperforms the previous network for the typical observational uncertainty range and maintains high accuracy even when subject to large uncertainties.

Reconstructing large-scale temperature profiles around z ∼ 6 quasars

ABSTRACT High-redshift quasars ionize He ii into He iii around them, heating the intergalactic medium in the process and creating large regions with elevated temperature. In this work, we demonstrate a method based on a convolutional neural network (CNN) to recover the spatial profile for T0, the temperature at the mean cosmic density, in quasar proximity zones. We train the neural network with synthetic spectra drawn from a Cosmic Reionization on Computers simulation. We discover that the simple CNN is able to recover the temperature profile with an accuracy of ≈1400 K in an idealized case of negligible observational uncertainties. We test the robustness of the CNN and discover that it is robust against the uncertainties in quasar host halo mass, quasar continuum, and ionizing flux. We also find that the CNN has good generality with regard to the hardness of quasar spectra. This shows that with noiseless spectra, one could use a simple CNN to distinguish gas inside or outside the He iii region created by the quasar. Because the size of the He iii region is closely related to the total quasar lifetime, this method has great potential in constraining the quasar lifetime on ∼Myr time-scales. However, noise poses a big problem for accuracy and could downgrade the accuracy to ≈2340 K even for very high signal-to-noise (≳50) spectra. Future studies are needed to reduce the error associated with noise to constrain the lifetimes of reionization epoch quasars with currently available data.

Reconciling different approaches to quantifying land surface temperature impacts of afforestation using satellite observations

Abstract. Satellite observations have been widely used to examine afforestation effects on local surface temperature at large spatial scales. Different approaches, which potentially lead to differing definitions of the afforestation effect, have been used in previous studies. Despite their large differences, the results of these studies have been used in climate model validation and cited in climate synthesis reports. Such differences have been simply treated as observational uncertainty, which can be an order of magnitude bigger than the signal itself. Although the fraction of the satellite pixel actually afforested has been noted to influence the magnitude of the afforestation effect, it remains unknown whether it is a key factor which can reconcile the different approaches. Here, we provide a synthesis of three influential approaches (one estimates the actual effect and the other two the potential effect) and use large-scale afforestation over China as a test case to examine whether the different approaches can be reconciled. We found that the actual effect (ΔTa) often relates to incomplete afforestation over a medium-resolution satellite pixel (1 km). ΔTa increased with the afforestation fraction, which explained 89 % of its variation. One potential effect approach quantifies the impact of quasi-full afforestation (ΔTp1), whereas the other quantifies the potential impact of full afforestation (ΔTp2) by assuming a shift from 100 % openland to 100 % forest coverage. An initial paired-sample t test shows that ΔTa<ΔTp1<ΔTp2 for the cooling effect of afforestation ranging from 0.07 to 1.16 K. But when all three methods are normalized for full afforestation, the observed range in surface cooling becomes much smaller (0.79 to 1.16 K). Potential cooling effects have a value in academic studies where they can be used to establish an envelope of effects, but their realization at large scales is challenging given its nature of scale dependency. The reconciliation of the different approaches demonstrated in this study highlights the fact that the afforestation fraction should be accounted for in order to bridge different estimates of surface cooling effects in policy evaluation.

Bayes goes fast: Uncertainty quantification for a covariant energy density functional emulated by the reduced basis method

A covariant energy density functional is calibrated using a principled Bayesian statistical framework informed by experimental binding energies and charge radii of several magic and semi-magic nuclei. The Bayesian sampling required for the calibration is enabled by the emulation of the high-fidelity model through the implementation of a reduced basis method (RBM)—a set of dimensionality reduction techniques that can speed up demanding calculations involving partial differential equations by several orders of magnitude. The RBM emulator we build—using only 100 evaluations of the high-fidelity model—is able to accurately reproduce the model calculations in tens of milliseconds on a personal computer, an increase in speed of nearly a factor of 3,300 when compared to the original solver. Besides the analysis of the posterior distribution of parameters, we present model calculations for masses and radii with properly estimated uncertainties. We also analyze the model correlation between the slope of the symmetry energy L and the neutron skin of 48Ca and 208Pb. The straightforward implementation and outstanding performance of the RBM makes it an ideal tool for assisting the nuclear theory community in providing reliable estimates with properly quantified uncertainties of physical observables. Such uncertainty quantification tools will become essential given the expected abundance of data from the recently inaugurated and future experimental and observational facilities.

Robust Lane Change Decision Making for Autonomous Vehicles: An Observation Adversarial Reinforcement Learning Approach

Reinforcementlearning holds the promise of allowing autonomous vehicles to learn complex decision making behaviors through interacting with other traffic participants. However, many real-world driving tasks involve unpredictable perception errors or measurement noises which may mislead an autonomous vehicle into making unsafe decisions, even cause catastrophic failures. In light of these risks, to ensure safety under perception uncertainty, autonomous vehicles are required to be able to cope with the worst case observation perturbations. Therefore, this paper proposes a novel observation adversarial reinforcement learning approach for robust lane change decision making of autonomous vehicles. A constrained observation-robust Markov decision process is presented to model lane change decision making behaviors of autonomous vehicles under policy constraints and observation uncertainties. Meanwhile, a black-box attack technique based on Bayesian optimization is implemented to approximate the optimal adversarial observation perturbations efficiently. Furthermore, a constrained observation-robust actor-critic algorithm is advanced to optimize autonomous driving lane change policies while keeping the variations of the policies attacked by the optimal adversarial observation perturbations within bounds. Finally, the robust lane change decision making approach is evaluated in three stochastic mixed traffic flows based on different densities. The results demonstrate that the proposed method can not only enhance the performance of an autonomous vehicle but also improve the robustness of lane change policies against adversarial observation perturbations.

Templates of expected measurement uncertainties for total neutron cross-section observables

This paper provides a template of expected uncertainties and correlations for measurements of total neutron cross-section observables by transmission. Measurements with time-of-flight and mono-energetic neutron sources are covered. The information required for evaluations in the resonance region and high energy region is detailed, along with the template of uncertainties and correlations that can be used in the absence of other information.

On the Attribution of Mercury’s Perihelion Precession

Although Newtonian gravity and general relativity predicted the precession of Mercury perihelion historically, many improved methods continue to predict the precession of Mercury during recent decades of years. Uncertainties in various predictions and observations suggest that the attribution of Mercury’s precession is still not well understood. This paper argues that the cause of Mercury’s precession is not gravity, but the inertia of material motion left over from the formation of the solar system. According to this inertia theory, the planetary precession is associated with the ratio of total mass-energy density of the system to the mass-energy of the Sun and its change over time. If other factors are not changed with time, the perihelion precession of planets per orbit is proportional to his distance relative to the Sun. The conclusions of this paper can provide more effective factor considerations for the complete description of various astronomical events and phenomena using general relativity equations.

Templates of expected measurement uncertainties for neutron-induced capture and charged-particle production cross section observables

This paper provides a template of expected uncertainties and correlations for measurements of neutron-induced capture and charged-particle production cross sections. Measurements performed in-beam include total absorption spectroscopy, total energy detection,γ-ray spectroscopy, and direct charged-particle detection. Offline measurements include activation analysis and accelerator mass spectrometry. The information needed for proper use of the datasets in resonance region and high energy region evaluations is described, and recommended uncertainties are provided when specific values are not available for a dataset.

Multistate model to estimate sex‐specific dispersal rates and distances for a wetland‐breeding amphibian population

AbstractHow animals move across space and time in heterogeneous landscapes has important implications for the conservation of imperiled and dispersal‐limited taxa, such as amphibians. Wetland‐breeding amphibians are thought to exhibit strong site fidelity, but there is growing evidence that many species are more vagile than previously assumed. Few studies have quantified breeding dispersal probabilities or distances while also accounting for observational uncertainty, which means that resultant estimates could be biased or confounded with sampling intensity and other demographic rates (e.g., site‐ and time‐specific detection and survival probabilities). Here, we analyze data from a 6‐year capture–mark–recapture study on adult spotted salamanders (Ambystoma maculatum) conducted at 12 wetlands in central Pennsylvania to estimate population dynamics and population structure at multiple scales. We used a multistate, hidden Markov estimator to quantify sex‐specific site fidelity and breeding dispersal as a function of Euclidean distance between wetlands while accounting for imperfect detection. We estimated short‐timescale movements (i.e., those that occur within a single breeding season) and longer timescale movements (i.e., those that occur among breeding seasons) to determine if dispersal rates and distances differed. We found that interannual site fidelity of males varied among wetlands and was positively associated with population density. Females exhibited higher interannual site fidelity and dispersed further than males between breeding seasons. Within breeding seasons, we found that up to 6% of males dispersed to a new wetland each day. Our study is the first to directly compare amphibian breeding dispersal probabilities and distances at multiple scales and provides a robust framework for improving inference on the spatial and temporal patterns of amphibian movement.

Unreliable networks with random parameter matrices and time-correlated noises: distributed estimation under deception attacks.

This paper examines the distributed filtering and fixed-point smoothing problems for networked systems, considering random parameter matrices, time-correlated additive noises and random deception attacks. The proposed distributed estimation algorithms consist of two stages: the first stage creates intermediate estimators based on local and adjacent node measurements, while the second stage combines the intermediate estimators from neighboring sensors using least-squares matrix-weighted linear combinations. The major contributions and challenges lie in simultaneously considering various network-induced phenomena and providing a unified framework for systems with incomplete information. The algorithms are designed without specific structure assumptions and use a covariance-based estimation technique, which does not require knowledge of the evolution model of the signal being estimated. A numerical experiment demonstrates the applicability and effectiveness of the proposed algorithms, highlighting the impact of observation uncertainties and deception attacks on estimation accuracy.

Verification and estimation of uncertainties of Tobias Mayer's 18th century astronomical observations

Perceiving the uncertainty of the measurement has been changing over the past centuries, reflecting the advancement in the experimental techniques, the urge for reliable and reproducible measurement methodology, and development of mathematical data processing and evaluating algorithms. From the historical perspective, the concepts of considering the measurement uncertainty were firstly introduced with geographic and cartographic measurements. In this context, the works of Tobias Mayer on lunar landscape measurements are widely highlighted which, at that time, presented innovative approaches in data processing with the method of averages and pioneeringly addressed the issue of measurement error. In this study, we analyze in details the Mayer’s set of 27 non-linear equations with 3 unknown parameters and discuss the effect of Mayer’s linearization and subsequent mathematical procedures on the accuracy of the parameter values in contrast with the results from rigorous treatment of non-linear regression model involving the least-square method. In particular, we compare the values of the unknown parameters and their uncertainties in several variants in the linearized and non-linearized model, providing monitoring of a small deviation of the Mayer’s linearization. The results, presented here, show that despite the conceptual and computational simplification of the Mayer’s method, such an approach to data processing can be exploited, with an acceptable level of accuracy, in several practical situations even today.

A decision support system based on Bayesian modelling for pest management: Application to wireworm risk assessment in maize fields

Protecting crops against pests is a major issue in the current agricultural production system. In particular, assessing the risk to crops can promote integrated pest management (IPM) strategies that encourage natural control mechanisms and advocate the use of pesticides as a last resort. In this study, we focused on wireworms, major soil-dwelling insect pests inflicting severe economic damage on various crops (including maize, potatoes and cereals) across Europe and North America. We have developed an original hierarchical Bayesian model that explicitly accounts for biological knowledge and uncertainty in field observations, rather than relying solely on statistical correlations, to predict the level of wireworm infestation. The model was calibrated and validated using a substantial dataset originating from an agro-environmental survey carried out over three consecutive years (2012–2014) in France, which provides the wireworm abundance in 419 maize fields, together with information on the landscape context, field history, weather conditions, soil characteristics and farming practices associated to each field. Model outcomes show good agreement with current knowledge from literature and field expertise in terms of the effects of variables on wireworm abundance, and provide fairly good predictive capacity. Subsequently, the model was encapsulated as a software (R shiny application) to predict the risk of wireworm infestation in any field of interest, and can be used by farmers or agricultural advisors as a decision support system for the implementation of IPM strategies. The conceptual framework that we implemented can be adapted to a wide range of similar situations involving other crops and pests.

Characterizing model uncertainty of upper-bound limit analysis on slopes using 3D rotational failure mechanism

The slope stability assessment is a classical problem in geotechnical engineering. This topic have attracted many researcher’s attention and various theoretical models for predicting critical slope heights or safety factors in the light of the limit equilibrium (LE) method and the kinematical approach of limit analysis (LA) method. Meanwhile, a large number of experimental studies have been conducted to check the slope stability. Using centrifuge testing results, this paper aims to employ Bayesian method to characterize the model uncertainties of the classical three-dimensional rotational failure mechanism proposed by Michalowski and Drescher (2009) to predict critical slope heights in frictional soils, by incorporating the test uncertainties and parameter uncertainties. The obtained results show that the LA three-dimensional rotational failure mechanism overestimates the critical slope height compared with the LE method, and the experimental observational uncertainty has negligible influences on the posterior statistics of model uncertainty.

Estimation of lidar-based gridded DEM uncertainty with varying terrain roughness and point density

Light detection and ranging (lidar) scanning systems can be used to provide a point cloud with high quality and point density. Gridded digital elevation models (DEMs) interpolated from laser scanning point clouds are widely used due to their convenience, however, DEM uncertainty is rarely provided. This paper proposes an end-to-end workflow to quantify the uncertainty (i.e., standard deviation) of a gridded lidar-derived DEM. A benefit of the proposed approach is that it does not require independent validation data measured by alternative means. The input point cloud requires per point uncertainty which is derived from lidar system observational uncertainty. The propagated uncertainty caused by interpolation is then derived by the general law of propagation of variances (GLOPOV) with simultaneous consideration of both horizontal and vertical point cloud uncertainties. Finally, the interpolated uncertainty is then scaled by point density and a measure of terrain roughness to arrive at the final gridded DEM uncertainty. The proposed approach is tested with two lidar datasets measured in Waikoloa, Hawaii, and Sitka, Alaska. Triangulated irregular network (TIN) interpolation is chosen as the representative gridding approach. The results indicate estimated terrain roughness/point density scale factors ranging between 1 (in flat areas) and 7.6 (in high roughness areas), with a mean value of 2.3 for the Waikoloa dataset and between 1 and 9.2 with a mean value of 1.2 for the Sitka dataset. As a result, the final gridded DEM uncertainties are estimated between 0.059 m and 0.677 m with a mean value of 0.164 m for the Waikoloa dataset and between 0.059 m and 1.723 m with a mean value of 0.097 m for the Sitka dataset.

Bayesian inference of real-time dynamics from lattice QCD

The computation of dynamical properties of nuclear matter, ranging from parton distribution functions of nucleons and nuclei to transport properties in the quark-gluon plasma, constitutes a central goal of modern theoretical physics. This real-time physics often defies a perturbative treatment and the most successful strategy so far is to deploy lattice QCD simulations. These numerical computations are based on Monte-Carlo sampling and formulated in an artificial Euclidean time. Real-time physics is most conveniently formulated in terms of spectral functions, which are hidden in lattice QCD behind an ill-posed inverse problem. I will discuss state-of-the art methods in the extraction of spectral functions from lattice QCD simulations, based on Bayesian inference and emphasize the importance of prior domain knowledge, vital to regularizing the otherwise ill-posed extraction task. With Bayesian inference allowing us to make explicit the uncertainty in both observations and in our prior knowledge, a systematic estimation of the total uncertainties in the extracted spectral functions is nowadays possible. Two implementations of the Bayesian Reconstruction (BR) method for spectral function extraction, one for MAP point estimates and one based on an open access Monte-Carlo sampler are provided. I will briefly touch on the use of machine learning for spectral function reconstruction and discuss some new insight it has brought to the Bayesian community.

TDOA positioning in the presence of atmospheric refraction and observer uncertainty influence analysis and correction

Due to the inhomogeneity of the atmosphere, radio waves are subjected to refraction during the propagation process, which reduces its propagation speed and bends the propagation path. If the influence is not considered, a large error will be caused by using time difference of arrival (TDOA). The influence of atmospheric refraction on the TDOA localization under the known elevation constraint is analyzed, a subsection iterative method is proposed to correct the localization error caused by atmospheric refraction, the Cramer-Rao lower bound (CRLB) for location estimation under atmospheric refraction is deduced. The effectiveness of the proposed method is validated through simulation results and analysis.

Learning and ensemble based MPC with differential dynamic programming for nuclear power autonomous control

Reactor autonomous control is one of the key research fields of advanced reactor R&D. Although some basic achievements have been made, it is still an unfinished goal to realize the intelligent autonomous control of the reactor. In this paper, some learning-based and ensemble-based predictive models and some model predictive control strategies combined with ensemble Kalman filter (EnKF) are proposed to realize a feasible path to the reactor autonomous control. For a typical point reactor and the operation target, reactor state space predictions are completed by mechanism-based models and data-driven learning-based models respectively. Mechanism models based on parameter perturbation and the neural network models that have higher accuracy are selected as predictors of model predictive control (MPC). The MPC results show that the model with less precision (e.g. support vector machine and random forest regression) has large oscillation results, while the predictive control based on neural network and mechanism model can effectively achieve the control objectives. To improve the accuracy of MPC, ensemble learning methodologies with a second learner based on linear regression, stacking different predictive models into one model are proposed. The accuracy of MPC with the ensemble model can exceed or be close to the optimal individual results. EnKF is proposed to estimate core state and covariance, which would be helpful to restrain the influence of observation uncertainties and predictive model error. In addition, some limitation of the proposed methodologies is discussed and needed further study.