Uncertainty In Reconstruction Research Articles

Seasonal Mg/Ca-inferred temperatures of brackish water ostracods

The common brackish water ostracods Cyprideis torosa and Loxoconcha elliptica frequently occur together in high abundances in marginal marine environments. Seasonality of calcification differs between species and can have important implications for palaeotemperature reconstructions. There are existing palaeotemperature calibrations for both genera. However, the Loxoconcha spp. Mg/Ca temperature calibration has not thus far been applied to L. elliptica. The equation for Loxoconcha spp. does not rely on a known Mg/Cawater value, unlike the calibration for C. torosa, suggesting it may be possible to reconstruct temperatures without an estimation of Mg/Cawater, which is potentially particularly beneficial in environments that have highly dynamic Mg/Cawater. However, the calibration has only been applied in environments with marine-like Mg/Cawater. Demonstrating the applicability of the equation in marginal marine environments (with Mg/Cawater 3–5 mol/mol) and tracking the seasonal calcification of L. elliptica alongside C. torosa, therefore, has the potential to improve uncertainty in seasonal palaeotemperature reconstructions. Here, we compare previous monitoring of C. torosa with L. elliptica from the same collections. We demonstrate that the Mg/Ca temperature calibration for Loxoconcha spp. is appropriate to use with L. elliptica. Mg/Ca-inferred temperatures broadly track spring temperatures and suggest spring calcification. Cyprideis torosa Mg/Ca-inferred temperatures record the range of expected temperatures between spring and autumn. When analysing multiple single valves of L. elliptica and C. torosa simultaneously, the maximum Mg/CaC.torosa can, therefore, be used to reconstruct maximum summer temperatures, the minimum Mg/CaC.torosa to reconstruct autumn temperatures, and the Mg/CaL.elliptica to reconstruct the range in spring temperatures.

Compressive sensing-based tomography for Absolute X-ray UltraViolet (AXUV) diagnostics.

Absolute x-ray ultraviolet diagnostics ensures 2D coverage of the radiation emission region that enables tomographic reconstruction. However, retrieving the local emissivity via tomography remains a challenge due to its ill-posed nature. Tikhonov regularization with smoothness operation generally performs well but tends to over-smooth regions with steep gradients and local structure in the radiation profile and may introduce artifacts. In this paper, a tomography method based on compressive sensing theory with Tikhonov regularization terms is developed. Experimental results on multiple phantom sets show that the proposed method improves the reconstruction accuracy and quality in regions with steep gradients compared with the Tikhonov regularization method and suppresses the unphysical negative emissivity. The analysis of reconstruction uncertainty shows that the dictionary learning process provides more accurate prior information about steep gradients to improve the quality of reconstructed images, and compressive sensing has the denoising capability to reduce the impact of noise. Finally, the method is validated by data from the Sino-UNIted Spherical Tokamak, showing fewer artifacts and more reliable reconstruction images than the earlier method.

Ensemble Modeling of Ice Sheet Dynamics in the Last Glacial Cycle

Ensemble simulations (taking into account to the uncertainty of paleoclimate reconstructions) with a model of ice sheet dynamics for the last glacial cycle (128 kyr) are carried out. In general, the model realistically reproduces the spatial structure of the ice sheets and the heights of their domes in the Northern Hemisphere, as well as the associated changes in the ocean level. Perturbations with a sufficiently large amplitude of paleoclimate data in the model show significant differences in the results of modeling the ice sheets of the Northern Hemisphere from the data obtained for the initial paleoreconstruction, including for the Last Glacial Maximum and the time interval of 58‒51 ka (the initial part of MIS3). According to the simulation results, the uncertainty of global reconstructions for the Last Glacial Maximum is 2°С, which is consistent with the existing estimates.

ENSEMBLE MODELLING OF ICE SHEET DYNAMICS IN THE LAST GLACIAL CYCLE

Ensemble simulations (taking into account uncertainty of paleoclimate reconstructions) with a models for ice sheets dynamics for the last glacial cycle (last 128 kyr) are carried out. The model realistically reproduces spatial structure of major ice sheets and heights of their domes in the Northern hemisphere as well as the associated changes in global sea level. Perturbations with a sufficiently large amplitude applied to the initial paleoreconstruction result in marked differences of the modelling, in particular, durimg the Last Glacial maximu and during 58–51 kyr before present (the initial part of MIS3). According to our simualtions, the uncertainty of temperature reconstructions durng the last galcial cycle is limited to 2°C in agreement with existing estimates.

Trends and variability in the Southern Annular Mode over the Common Era

The Southern Annular Mode (SAM) is the leading mode of atmospheric variability in the extratropical Southern Hemisphere and has wide ranging effects on ecosystems and societies. Despite the SAM’s importance, paleoclimate reconstructions disagree on its variability and trends over the Common Era, which may be linked to variability in SAM teleconnections and the influence of specific proxies. Here, we use data assimilation with a multi-model prior to reconstruct the SAM over the last 2000 years using temperature and drought-sensitive climate proxies. Our method does not assume a stationary relationship between the SAM and the proxy records and allows us to identify critical paleoclimate records and quantify reconstruction uncertainty through time. We find no evidence for a forced response in SAM variability prior to the 20th century. We do find the modern positive trend falls outside the 2σ range of the prior 2000 years at multidecadal time scales, supporting the inference that the SAM’s positive trend over the last several decades is a response to anthropogenic climate change.

LatentPCN: latent space-constrained point cloud network for reconstruction of 3D patient-specific bone surface models from calibrated biplanar X-ray images.

Accurate three-dimensional (3D) models play crucial roles in computer assisted planning and interventions. MR or CT images are frequently used to derive 3D models but have the disadvantages that they are expensive or involving ionizing radiation (e.g., CT acquisition). An alternative method based on calibrated 2D biplanar X-ray images is highly desired. A point cloud network, referred as LatentPCN, is developed for reconstruction of 3D surface models from calibrated biplanar X-ray images. LatentPCN consists of three components: an encoder, a predictor, and a decoder. During training, a latent space is learned to represent shape features. After training, LatentPCN maps sparse silhouettes generated from 2D images to a latent representation, which is taken as the input to the decoder to derive a 3D bone surface model. Additionally, LatentPCN allows for estimation of a patient-specific reconstruction uncertainty. We designed and conducted comprehensive experiments on datasets of 25 simulated cases and 10 cadaveric cases to evaluate the performance of LatentLCN. On these two datasets, the mean reconstruction errors achieved by LatentLCN were 0.83mm and 0.92mm, respectively. A correlation between large reconstruction errors and high uncertainty in the reconstruction results was observed. LatentPCN can reconstruct patient-specific 3D surface models from calibrated 2D biplanar X-ray images with high accuracy and uncertainty estimation. The sub-millimeter reconstruction accuracy on cadaveric cases demonstrates its potential for surgical navigation applications.

Real-Time Neural Dense Elevation Mapping for Urban Terrain With Uncertainty Estimations

Having good knowledge of terrain information is essential for improving the performance of various downstream tasks on complex terrains, especially for the locomotion and navigation of legged robots. We present a novel framework for neural urban terrain reconstruction with uncertainty estimations. It generates dense robot-centric elevation maps online from sparse LiDAR observations. We design a novel pre-processing and point features representation approach that ensures high robustness and computational efficiency when integrating multiple point cloud frames. A generative Bayesian model then recovers the detailed terrain structures while simultaneously providing the pixel-wise reconstruction uncertainty. We evaluate the proposed pipeline through both simulation and real-world experiments. Our approach achieves high-quality terrain reconstruction with real-time performance on a mobile platform, and the uncertainty estimates may further benefit the downstream tasks of legged robots.

Resolving Uncertainties in Foraminifera-Based Relative Sea-Level Reconstruction: a Case Study from Southern New Zealand

Abstract Since the pioneering work of David Scott and others in the 1970s and 1980s, foraminifera have been used to develop precise sea-level reconstructions from salt marshes around the world. In New Zealand, reconstructions feature rapid rates of sea-level rise during the early to mid-20th century. Here, we test whether infaunality, taphonomy, and sediment compaction influence these reconstructions. We find that surface (0–1 cm) and subsurface (3–4 cm) foraminiferal assemblages show a high degree of similarity. A landward shift in assemblage zones is consistent with recent sea-level rise and transgression. Changes associated with infaunality and taphonomy do not affect transfer function-based sea-level reconstructions. Applying a geotechnical modelling approach to the core from which sea-level changes were reconstructed, we demonstrate compaction is also negligible, resulting in maximum post-depositional lowering of 2.5 mm. We conclude that salt-marsh foraminifera are indeed highly accurate and precise indicators of past sea levels.

Uncertainty Quantification in Intelligent-Based Electrical Resistivity Tomography Image Reconstruction With Monte Carlo Dropout Strategy

The Electrical Resistivity Tomography (ERT) method plays an essential role in researching electrical properties at the core scale, and the resistivity image inversion reconstruction technique is the key to constraining measurement accuracy. In recent years, with the improvement of computer operation capabilities and the acquisition of a large amount of geophysical data, the inversion algorithms represented by machine learning (ML) have made it possible to reconstruct ERT images automatically. However, the solution of ERT image inversion has large uncertainty due to high nonlinearity. Meanwhile, traditional ML methods are not initially developed for evaluating the uncertainty of such reconstruction results. In this paper, we propose a novel MC-Net ML scheme to quantitatively estimate the uncertainty of the ERT image reconstruction by introducing the Monte-Carlo Dropout strategy with a multiple stochastic method to approximate Bayesian inference. As one important evaluation index, the correlation coefficient (CC) between target image and reconstructed image is used to compare the effects of six types of simulated data carried by different ML schemes. The inversion results show that MC-Net increases lower limit of CC on the premise of ensuring average CC. Moreover, the accuracy of regression and classification results defined in this paper also obtained the highest value by MC-Net. With the proposed scheme, we can further provide a quantification of the uncertainty to determine whether reconstructed images are reliable after reconstructing the ERT images, which is of great significance for actual intelligent inversion tasks.

Neural Network-Based Adaptive Asymptotic Prescribed Performance Tracking Control of Hydraulic Manipulators

Hydraulic manipulators are extremely complicated systems due to the highly nonlinear characteristics, strong coupling among multiple joints, and heavy modeling uncertainties, which greatly complicate the high-performance tracking controller development compared with conventional manipulators driven by electrical motors. To overcome the above obstacles, this article proposes a novel radial basis function neural network (RBFNN)-based adaptive asymptotic prescribed performance controller for <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$n$ </tex-math></inline-formula> -degrees of freedom (DOF) hydraulic manipulators. First, the entire manipulator system model containing actuator dynamics is derived. Then, the adaptive asymptotic prescribed performance controller is synthesized based on the backstepping framework, in which the RBFNN is employed to estimate the unknown joint coupling dynamics, while the RBFNN reconstruction error and uncertainties of the actuator dynamics are handled by the robust integral of the sign of the error (RISE) feedbacks. Meanwhile, a prescribed performance function (PPF), which characterizes both the certain transient and steady-state performance, is innovatively incorporated into the control design to restrict the joint tracking errors. The theoretical analysis proves that the proposed control strategy can achieve a prescribed tracking performance and the asymptotic stability of the whole closed-loop system can also be ensured. Finally, comparative simulations are conducted to verify the validity of the proposed controller.

Reconstructing mass and enrichment of uranium shell based on associated particle imaging

In the field of arms control verification, the attributes of uranium shells often need to be measured and authenticated. In this paper, the feasibility and results of reconstructing the masses and enrichment of uranium shells based on the associated particle imaging are studied, aiming at constructing and testifying the theoretical framework of this method. First, the basic principles of this method are introduced. Second, the physical model of the neutron transmission imaging of uranium shells was constructed, and the feasibility of reconstructing the masses of uranium shells based on the neutron transmission imaging was proved with the uncertainties of the reconstruction results analyzed. Third, the physical model of the neutron induced fission imaging of uranium shells was constructed, and reconstructing the enrichment of uranium shells based on the neutron transmission imaging and neutron induced fission imaging was proved to be feasible, and the influencing factors of the uncertainties of the reconstruction results were analyzed. Meanwhile, a method of reconstructing the enrichment of uranium shells based on a single calibration line and a method of revising neutron attenuation and multiplication effect are proposed; based on the above two methods, the reconstruction uncertainties of the enrichment of uranium shells can reach the level of 3.78%.

Localization of magnetocardiographic sources for myocardial infarction cases using deterministic and Bayesian approaches

In this paper, the inverse problems of cardiac sources using analytical and probabilistic methods are solved and discussed. The standard Tikhonov regularization technique is solved initially to estimate the under-determined heart surface potentials from Magnetocardiographic (MCG) signals. The results of the deterministic method subjected to noise in the measurements are discussed and compared with the probabilistic models. Hierarchical Bayesian modeling with fixed Gaussian prior is employed to quantify the uncertainties in source reconstructions. A novel application of Variational Bayesian inference approach has been presented to estimate the heart sources. The reconstruction results of Variational Bayesian model with non-stationary priors are compared with solutions of simplistic Bayesian approach; and the performances are evaluated using Root Mean Square Error (RMSE) and correlation co-efficient metrics. The Bayesian solutions in the study are also extended to localize the MCG sources for two types of Myocardial infarction cases.

Time Series Construction of Oregon and Washington Snowfall since 1890 and an Update of California Snowfall through 2020

Abstract Time series of snowfall observations from over 500 stations in Oregon (OR) and Washington (WA) were generated for subregions of these states. Data problems encountered were as follows: 1) monthly totals in printed reports prior to 1940 that were not in the digital archive, 2) archived data listed as “missing” that were available, 3) digitized reports after 2010 eliminated good data, and 4) “zero” totals incorrectly entered in the official archive rather than “missing,” especially after 1980. Though addressing these was done, there is reduced confidence that some regional time series are representative of true long-term trends, especially for regions with few systematically reporting stations. For most regions characterized by consistent monitoring and with the most robust statistical reproducibility, we find no statistically significant trends in their periods of record (up to 131 years) for November–April seasonal totals through April 2020. This result includes the main snowfall regions of the Cascade Range. However, snowfall in some lower-elevation areas of OR and WA appear to have experienced declining trends, consistent with an increase in northeastern Pacific Ocean temperatures. Finally, previously constructed time series through April 2011 for regions in California are updated through April 2020 to include the recent, exceptionally low seasonal totals on the western slopes of the Sierra Nevada. This update indicates 2014/15 was the record lowest, 2013/14 was the 5th lowest, and 2012/13 was the 14th lowest of 142 years. Even so, the 1879–2020 linear trend in this key watershed region, though −2.6% decade−1, was not significantly different from zero due to high interannual variability and reconstruction uncertainty.

Empirical evidence for multidecadal scale global atmospheric electric circuit modulation by the El Niño-Southern Oscillation

The El Niño-Southern Oscillation (ENSO) modifies precipitation patterns across the planet. Charge separation in disturbed weather and thunderstorms drives the global atmospheric electric circuit (GEC), hence ENSO-induced precipitation changes are anticipated to affect the global circuit. By analysing historical atmospheric electricity data using a new data processing procedure based on the Carnegie curve, signals correlated with ENSO sea surface temperature (SST) anomalies are revealed. These demonstrate a persistent ENSO-GEC relationship for the majority of the twentieth century, in potential gradient data from Lerwick, Shetland and Watheroo, W. Australia. The recovered data is weighted towards the first half of the UTC day, giving a GEC sensitivity up to ∼5% °C−1 of SST anomaly in the Niño 3.4 and 4 regions of the Pacific Ocean. Transferring ENSO variability by electrical means represents an unexplored teleconnection, for example, through proposed GEC effects on stratiform cloud microphysics. The strong ENSO-GEC relationship also provides a quality test for historical atmospheric electricity data, and encourages their use in reducing SST reconstruction uncertainties.

Learning Reconstructability for Drone Aerial Path Planning

We introduce the first learning-based reconstructability predictor to improve view and path planning for large-scale 3D urban scene acquisition using unmanned drones. In contrast to previous heuristic approaches, our method learns a model that explicitly predicts how well a 3D urban scene will be reconstructed from a set of viewpoints. To make such a model trainable and simultaneously applicable to drone path planning, we simulate the proxy-based 3D scene reconstruction during training to set up the prediction. Specifically, the neural network we design is trained to predict the scene reconstructability as a function of the proxy geometry , a set of viewpoints, and optionally a series of scene images acquired in flight. To reconstruct a new urban scene, we first build the 3D scene proxy, then rely on the predicted reconstruction quality and uncertainty measures by our network, based off of the proxy geometry, to guide the drone path planning. We demonstrate that our data-driven reconstructability predictions are more closely correlated to the true reconstruction quality than prior heuristic measures. Further, our learned predictor can be easily integrated into existing path planners to yield improvements. Finally, we devise a new iterative view planning framework, based on the learned reconstructability, and show superior performance of the new planner when reconstructing both synthetic and real scenes.

Internal climate variability and spatial temperature correlations during the past 2000 years

Abstract. The spatio-temporal structure of natural climate variability has to be taken into account when unravelling observed climatic changes and simulating future climate change. However, based on the comparison of temperature reconstructions and climate model simulations covering the past 2 millennia, it has been argued that climate models are biased. They would simulate too little temporal temperature variability and too high correlations between temperature time series from different continents. One of the proposed causes is the lack of internal climate variability in climate models on centennial timescales, for instance variability related to the Atlantic meridional overturning circulation (AMOC). We present a perturbed-parameter ensemble with the iLOVECLIM Earth system model containing various levels of AMOC-related internal climate variability to investigate the effect on the spatio-temporal temperature variability structure. The model ensemble shows that enhanced AMOC variability indeed leads to more continental-scale temperature variability, but it also increases the spatio-temporal temperature correlations between different continents. However, combining the iLOVECLIM results with CMIP5 model results and various PAGES-2k temperature field reconstructions, we show overall agreement for the magnitude of continental temperature variability in models and reconstructions, but both the simulated and the reconstructed ranges are large. This is even more true when considering higher-order metrics like inter-continental temperature correlations or temperature variability land–sea contrasts. For such metrics, uncertainties in both model results and temperature reconstructions are so large that they hamper our ability to constrain simulated spatio-temporal structure of centennial temperature variability. As a result, we cannot determine the importance of AMOC variability for the climatic evolution over the past 2 millennia.

Incorporating uncertainty bounds in daily deformable dose accumulation for adaptive radiation therapy of head-and-neck cancer.

The widespread use of deformable dose accumulation (DDA) in adaptive radiotherapy (ART) has been limited due to the lack of clinically compatible methods to consider its related uncertainties. We estimate dose reconstruction uncertainties in daily DDA during CT-guided radiotherapy of head-and-neck cancer (HNC). We project confidence intervals of cumulative dose-volume parameters to the parotids and determine threshold values to guide clinical decision-making in ART. Doses from daily images (megavoltage CTs [MVCTs]) of 20 HNC patients treated with tomotherapy were reconstructed and accumulated in the planning CT (PCT) utilizing a commercial DDA algorithm (PreciseART, Accuray, Inc.). For each mapped fraction, we warped the planning contours to the MVCT. Dose-volume histograms (DVHs) calculated in the MVCT (with warped contour and native dose) and the PCT (with native contour and mapped dose) were compared; the observed inconsistencies were associated with dose reconstruction errors. We derived uncertainty bounds for the transferred dose to voxels within the structure of interest in the PCT. The confidence intervals of cumulative dose-volume parameters were mid-treatment projected and evaluated as predictors of the end of treatment cumulative metrics. The need for plan adaptation was tested by comparing the projected uncertainty bounds with the treatment constraint points. Among all cases, the uncertainty in mean values of daily dose distributions mapped to the reference parotid's contours averaged between 2.8% and 3.8% of typical single fraction planning values and less than 1% for the planning target volume (PTV) D95%. These daily inconsistencies were higher in the ipsilateral compared to the contralateral parotid and increased toward the end of treatment. The magnitude of the uncertainty bounds for the cumulative treatment mean dose, D50%, and V20Gy to the parotids, and PTV D95% were on average 3.5%, 6.6%, 4.6%, and 0.4% of the planned or prescribed values, with confidence intervals of 97.1%-107.0%, 98.2%-110.4%, 95.6%-111.1%, and 98.2%-100.2% respectively. The uncertainty intervals projected at mid-treatment intersected with the end of treatment bounds in 82% of the parotid's metrics; half of them presented an overlapping percentage greater than 60%. In five patients, the cumulative mean doses were projected at mid-treatment to exceed the total treatment constraint point by at least 3%; this threshold was exceeded at the end of treatment in the five cases. Underdosing was projected in only one case; the cumulative PTV D95% at the end of treatment was below the clinical threshold. Uncertainty bounds were incorporated into the results of a commercial DDA tool. The cohort's statistics showed that the parotids' cumulative DVH metrics frequently exceeded the planning values if confidence intervals were included. Most of the uncertainty bounds of the PTV metrics were kept within the clinical thresholds. We verified that mid-treatment violation projections led to exceeding the constraint point at the end of the treatment. Based on a 3% threshold, approximately one fourth of the patients are expected to be replanned at mid-treatment for parotids sparing during HNC radiotherapy.

Improved winter data coverage of the Southern Ocean CO2 sink from extrapolation of summertime observations

The Southern Ocean is an important sink of anthropogenic CO2, but it is among the least well-observed ocean basins, and consequentially substantial uncertainties in the CO2 flux reconstruction exist. A recent attempt to address historically sparse wintertime sampling produced ‘pseudo’ wintertime observations of surface pCO2 using subsurface summertime observations south of the Antarctic Polar Front. Here, we present an estimate of the Southern Ocean CO2 sink that combines a machine learning-based mapping method with an updated set of pseudo observations that increases regional wintertime data coverage by 68% compared with the historical dataset. Our results confirm the suggestion that improved winter coverage has a modest impact on the reconstruction, slightly strengthening the uptake trend in the 2000s. After also adjusting for surface boundary layer temperature effects, we find a 2004-2018 mean sink of −0.16 ± 0.07 PgC yr−1 south of the Polar Front and −1.27 ± 0.23 PgC yr−1 south of 35°S, consistent with independent estimates from atmospheric data.

Iterative Construction of the Optimal Sunspot-Number Series

The relative number of sunspots represents the longest evidence describing the level of solar activity. As such, its use goes beyond solar physics, e.g. towards climate research. The construction of a single representative series is a delicate task which involves a combination of observation of many observers. We propose a new iterative algorithm that allows to construct a target series of relative sunspot number of a hypothetical stable observer by optimally combining series obtained by many observers. We show that our methodology provides us with results that are comparable with recent reconstructions of both sunspot number and group number. Furthermore, the methodology accounts for the possible non-solar changes of observers' time series such as gradually changing observing conditions or slow change in the observers vision. It also provides us with reconstruction uncertainties. We apply the methodology to a limited sample of observations by \v{C}ESLOPOL network and discuss its properties and limitations.

New paleoclimatic evidence of an extraordinary rise in temperature in the Northern Hemisphere in the last 3–4 decades

ABSTRACT Prognosis of temperature changes in the Northern Hemisphere for the period 1980–2020 was made using seven temperature paleoreconstructions covering the last 1–2 millennia and ending 1979–2016. Forecasts were made using the analogue nonlinear prediction method. A part of paleodata before 1800 (prior to the beginning of anthropogenic impact) was used as an information bank. In all seven cases, the forecast gave either a decrease or a very slight increase in temperature during 1980–2020. Statistical experiments performed with using prediction errors based on a conservative estimation of reconstruction uncertainties showed that a temperature increase of 0.25°C in the specified epoch is not excluded, but its probability is low (P < 0.20). This means that if the climate in the 20th - early 21st century was controlled by the same dynamic system as before 1800, the noticeable warming of the Northern Hemisphere in 1980–2020 should not be observed. Thus, it was shown that the data of modern paleoclimatology confirm that the climate of the Northern Hemisphere in the last 40–50 years was significantly influensed by an additional factor that did not act in the previous 1–2 millennia. It was also shown that if the actual uncertainties are significantly higher than the conservative estimates used, a warming of 0.5 degrees between 1980 and 2020 due to internal climate variability becomes possible.