Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Ensemble models”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 667 records · Page 37
All sourcesNASA NTRSDOE OSTI
NASA NTRS2021

Evaluation of Clouds, Radiation, and Precipitation in CMIP6 Models Using Global Weather States Derived from ISCCP-H Cloud Property Data

A clustering methodology is applied to cloud optical depth cloud top pressure (TAU-PC) histograms from the new, 1-degree resolution, ISCCP-H dataset, to derive an updated global Weather State (WS) dataset. Then, PC-TAU histograms from current-climate CMIP6 model simulations are assigned to the ISCCP-H WSs along with their concurrent radiation and precipitation properties, to evaluate model cloud, radiation, and precipitation properties in the context of the Weather States. The new ISCCP-H analysis produces WSs that are very similar to those previously found in the lower resolution ISCCP-D dataset. The main difference lies in the splitting of the ISCCP-D thin stratocumulus WS between the ISCCP-H shallow cumulus and stratocumulus WSs, which results in the reduction by one of the total WS number. The evaluation of the CMIP6 models against the ISCCP-H Weather States, shows that, in the ensemble mean, the models are producing an adequate representation of the frequency and geographical distribution of the WSs, with measurable improvements compared to the WSs derived for the CMIP5 ensemble. However, the frequency of shallow cumulus clouds continues to be underestimated, and, in some WSs the good agreement of the ensemble mean with observations comes from averaging models that significantly overpredict and underpredict the ISCCP-H WS frequency. In addition, significant biases exist in the internal cloud properties of the model WSs, such as the model underestimation of cloud fraction in middle-top clouds and secondarily in midlatitude storm and stratocumulus clouds, that result in an underestimation of cloud SW cooling in those regimes.

Clouds↗
NASA NTRS2013

On the Role of SST Forcing in the 2011 and 2012 Extreme U.S. Heat and Drought: A Study in Contrasts

This study compares the extreme heat and drought that developed over the United States in 2011 and 2012 with a focus on the role of SST forcing. Experiments with the NASA GEOS-5 atmospheric general circulation model show that the winter/spring response over the U.S. to the Pacific SST is remarkably similar for the two years despite substantial differences in the tropical Pacific SST. As such, the pronounced winter and early spring temperature differences between the two years (warmth confined to the south in 2011 and covering much of the continent in 2012) primarily reflect differences in the contributions from the Atlantic and Indian Oceans, with both acting to cool the east and upper mid-west during 2011, while during 2012 the Indian Ocean reinforced the Pacific-driven continental-wide warming and the Atlantic played a less important role. During late spring and summer of 2011 the tropical Pacific SST force a continued warming and drying over the southern U.S., though considerably weaker than observed. Nevertheless, the observed anomalies fall within the models intra-ensemble spread. In contrast, the rapid development of intense heat and drying over the central U.S. during June and July of 2012 falls outside the models intra-ensemble spread. The response to the SST (a northward expansion of a modest summer warming linked to the Atlantic) gives little indication that 2012 would produce record-breaking precipitation deficits and heat in the central Great Plains. A diagnosis of the 2012 observed circulation anomalies shows that the most extreme heat and drought was tied to the development of a stationary Rossby wave and an associated anomalous upper tropospheric high maintained by weather transients.

SST Forcing↗
NASA NTRS2009

Validation of Model Forecasts of the Ambient Solar Wind

Independent and automated validation is a vital step in the progression of models from the research community into operational forecasting use. In this paper we describe a program in development at the CCMC to provide just such a comprehensive validation for models of the ambient solar wind in the inner heliosphere. We have built upon previous efforts published in the community, sharpened their definitions, and completed a baseline study. We also provide first results from this program of the comparative performance of the MHD models available at the CCMC against that of the Wang-Sheeley-Arge (WSA) model. An important goal of this effort is to provide a consistent validation to all available models. Clearly exposing the relative strengths and weaknesses of the different models will enable forecasters to craft more reliable ensemble forecasting strategies. Models of the ambient solar wind are developing rapidly as a result of improvements in data supply, numerical techniques, and computing resources. It is anticipated that in the next five to ten years, the MHD based models will supplant semi-empirical potential based models such as the WSA model, as the best available forecast models. We anticipate that this validation effort will track this evolution and so assist policy makers in gauging the value of past and future investment in modeling support.

Macneice, P. J.↗
NASA NTRS2019

Investigating Sources of Variability and Error in Simulations of Carbon Dioxide in an Urban Region

Greenhouse gas (GHG) emissions estimation methods that use atmospheric trace gas observations, including inverse modeling techniques, perform better when carbon dioxide (CO2) fluxes are more accurately transported and dispersed in the atmosphere by a numerical model. In urban areas, transport and dispersion is particularly difficult to simulate using current mesoscale meteorological models due, in part, to added complexity from surface heterogeneity and fine spatial/temporal scales. It is generally assumed that the errors in GHG estimation methods in urban areas are dominated by errors in transport and dispersion. Other significant errors include, but are not limited to, those from assumed emissions magnitude and spatial distribution. To assess the predictability of simulated trace gas mole fractions in urban observing systems using a numerical weather prediction model, we employ an Eulerian model that combines traditional meteorological variables with multiple passive tracers of atmospheric CO2 from anthropogenic inventories and a biospheric model. The predictability of the Eulerian model is assessed by comparing simulated atmospheric CO2 mole fractions to observations from four in situ tower sites (three urban and one rural) in the Washington DC/Baltimore, MD area for February 2016. Four different gridded fossil fuel emissions inventories along with a biospheric flux model are used to create an ensemble of simulated atmospheric CO2 observations within the model. These ensembles help to evaluate whether the modeled observations are impacted more by the underlying emissions or transport. The spread of modeled observations using the four emission fields indicates the model's ability to distinguish between the different inventories under various meteorological conditions. Overall, the Eulerian model performs well; simulated and observed average CO2 mole fractions agree within 1% when averaged at the three urban sites across the month. However, there can be differences greater than 10% at any given hour, which are attributed to complex meteorological conditions rather than differences in the inventories themselves. On average, the mean absolute error of the simulated compared to actual observations is generally twice as large as the standard deviation of the modeled mole fractions across the four emission inventories. This result supports the assumption, in urban domains, that the predicted mole fraction error relative to observations is dominated by errors in model meteorology rather than errors in the underlying fluxes in winter months. As such, minimizing errors associated with atmospheric transport and dispersion may help improve the performance of GHG estimation models more so than improving flux priors in the winter months. We also find that the errors associated with atmospheric transport in urban domains are not restricted to certain times of day. This suggests that atmospheric inversions should use CO2 observations that have been filtered using meteorological observations rather than assuming that meteorological modeling is most accurate at certain times of day (such as using only mid-afternoon observations).

CO2↗
DOE OSTI2022

Generative Ensemble Regression: Learning Particle Dynamics from Observations of Ensembles with Physics-informed Deep Generative Models

Here, we propose a new method for inferring the governing stochastic ordinary differential equations (SODEs) by observing particle ensembles at discrete and sparse time instants, i.e., multiple “snapshots.” Particle coordinates at a single time instant, possibly noisy or truncated, are recorded in each snapshot but are unpaired across the snapshots. By training a physics-informed generative model that generates “fake” sample paths, we aim to fit the observed particle ensemble distributions with a curve in the probability measure space, which is induced from the inferred particle dynamics. We employ different metrics to quantify the differences between distributions, e.g., the sliced Wasserstein distances and the adversarial losses in generative adversarial networks. We refer to this method as generative “ensemble-regression” (GER), in analogy to the classic “point-regression,” where we infer the dynamics by performing regression in the Euclidean space. We illustrate the GER by learning the drift and diffusion terms of particle ensembles governed by SODEs with Brownian motions and Lévy processes up to 100 dimensions. We also discuss how to treat cases with noisy or truncated observations. Apart from systems consisting of independent particles, we also tackle nonlocal interacting particle systems with unknown interaction potential parameters by constructing a physics-informed loss function. Finally, we investigate scenarios of paired observations and discuss how to reduce the dimensionality in such cases by proving a convergence theorem that provides theoretical support.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

The future of the El Niño–Southern Oscillation: using large ensembles to illuminate time-varying responses and inter-model differences

Future changes in the El Niño–Southern Oscillation (ENSO) are uncertain, both because future projections differ between climate models and because the large internal variability of ENSO clouds the diagnosis of forced changes in observations and individual climate model simulations. By leveraging 14 single model initial-condition large ensembles (SMILEs), we robustly isolate the time-evolving response of ENSO sea surface temperature (SST) variability to anthropogenic forcing from internal variability in each SMILE. We find nonlinear changes in time in many models and considerable inter-model differences in projected changes in ENSO and the mean-state tropical Pacific zonal SST gradient. We demonstrate a linear relationship between the change in ENSO SST variability and the tropical Pacific zonal SST gradient, although forced changes in the tropical Pacific SST gradient often occur later in the 21st century than changes in ENSO SST variability, which can lead to departures from the linear relationship. Single-forcing SMILEs show a potential contribution of anthropogenic forcing (aerosols and greenhouse gases) to historical changes in ENSO SST variability, while the observed historical strengthening of the tropical Pacific SST gradient sits on the edge of the model spread for those models for which single-forcing SMILEs are available. Our results highlight the value of SMILEs for investigating time-dependent forced responses and inter-model differences in ENSO projections. The nonlinear changes in ENSO SST variability found in many models demonstrate the importance of characterizing this time-dependent behavior, as it implies that ENSO impacts may vary dramatically throughout the 21st century.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2021

Reduced El NiñO Variability in the Mid-Pliocene According to the Pliomip2 Ensemble

The mid-Pliocene warm period (3.264–3.025 Ma) is the most recent geological period during which atmospheric CO 2 levels were similar to recent historical values (∼400 ppm). Several proxy reconstructions for the mid-Pliocene show highly reduced zonal sea surface temperature (SST) gradients in the tropical Pacific Ocean, indicating an El Niño-like mean state. However, past modelling studies do not show these highly reduced gradients. Efforts to understand mid-Pliocene climate dynamics have led to the Pliocene Model Intercomparison Project (PlioMIP). Results from the first phase (PlioMIP1) showed clear El Niño variability (albeit significantly reduced) and did not show the greatly reduced time-mean zonal SST gradient suggested by some of the proxies. In this work, we study El Niño–Southern Oscillation (ENSO) variability in the PlioMIP2 ensemble, which consists of additional global coupled climate models and updated boundary conditions compared to PlioMIP1. We quantify ENSO amplitude, period, spatial structure and “flavour”, as well as the tropical Pacific annual mean state in mid-Pliocene and pre-industrial simulations. Results show a reduced ENSO amplitude in the model-ensemble mean (−24 %) with respect to the pre-industrial, with 15 out of 17 individual models showing such a reduction. Furthermore, the spectral power of this variability considerably decreases in the 3–4-year band. The spatial structure of the dominant empirical orthogonal function shows no particular change in the patterns of tropical Pacific variability in the model-ensemble mean, compared to the pre-industrial. Although the time-mean zonal SST gradient in the equatorial Pacific decreases for 14 out of 17 models (0.2 ∘C reduction in the ensemble mean), there does not seem to be a correlation with the decrease in ENSO amplitude. The models showing the most “El Niño-like” mean state changes show a similar ENSO amplitude to that in the pre-industrial reference, while models showing more “La Niña-like” mean state changes generally show a large reduction in ENSO variability. The PlioMIP2 results show a reasonable agreement with both time-mean proxies indicating a reduced zonal SST gradient and reconstructions indicating a reduced, or similar, ENSO variability.

mid-Pliocene warm period↗
DOE OSTI2026

New framework for benchmarking decadal predictions leveraging the PCMDI Metric Package with interactive visualization

Reliable climate predictions across multiple timescales are increasingly critical as climate-related risks continue to rise. With the growing number and diversity of climate prediction systems, systematic intercomparison has become essential. Here, we present a comprehensive evaluation framework based on the PCMDI Metric Package to assess the performance of multiple decadal climate prediction systems. Unlike uninitialized simulations, initialized predictions exhibit bias and predictive skill that evolve with forecast lead time. To address this, we introduce (1) model-by-lead-time portrait plots, which efficiently summarize metrics of global temperature, precipitation, and Arctic/Antarctic sea-ice extent, and (2) an HTML-based interactive visualization platform that provides detailed regional and seasonal diagnostics of model bias, skill scores, and ensemble spread for each model and lead time. Comparisons with uninitialized simulations further quantify the relative impacts of initialization and external forcing on prediction skill. The proposed framework provides a scalable and transparent approach for multi-model climate prediction assessments and can be readily extended to a wide range of operational and research forecasting systems.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Knowledge-informed deep learning for hydrological model calibration: an application to Coal Creek Watershed in Colorado

Abstract. Deep learning (DL)-assisted inverse mapping has shown promise in hydrological model calibration by directly estimating parameters from observations. However, the increasing computational demand for running the state-of-the-art hydrological model limits sufficient ensemble runs for its calibration. In this work, we present a novel knowledge-informed deep learning method that can efficiently conduct the calibration using a few hundred realizations. The method involves two steps. First, we determine decisive model parameters from a complete parameter set based on the mutual information (MI) between model responses and each parameter computed by a limited number of realizations (∼50). Second, we perform more ensemble runs (e.g., several hundred) to generate the training sets for the inverse mapping, which selects informative model responses for estimating each parameter using MI-based parameter sensitivity. We applied this new DL-based method to calibrate a process-based integrated hydrological model, the Advanced Terrestrial Simulator (ATS), at Coal Creek Watershed, CO. The calibration is performed against observed stream discharge (Q) and remotely sensed evapotranspiration (ET) from the water year 2017 to 2019. Preliminary MI analysis on 50 realizations resulted in a down-selection of 7 out of 14 ATS model parameters. Then, we performed a complete MI analysis on 396 realizations and constructed the inverse mapping from informative responses to each of the selected parameters using a deep neural network. Compared with calibration using observations covering all time steps, the new inverse mapping improves parameter estimations, thus enhancing the performance of ATS forward model runs. The Nash–Sutcliffe efficiency (NSE) of streamflow predictions increases from 0.53 to 0.8 when calibrating against Q alone. Using ET observations, on the other hand, does not show much improvement on the performance of ATS modeling mainly due to both the uncertainty of the remotely sensed product and the insufficient coverage of the model ET ensemble in capturing the observation. By using observed Q only, we further performed a multiyear analysis and show that Q is best simulated (NSE > 0.8) by including in the calibration the dry-year flow dynamics that show more sensitivity to subsurface characteristics than the other wet years. Moreover, when continuing the forward runs till the end of 2021, the calibrated models show similar simulation performances during this evaluation period as the calibration period, demonstrating the ability of the estimated parameters in capturing climate sensitivity. Our success highlights the importance of leveraging data-driven knowledge in DL-assisted hydrological model calibration.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Data-Efficient Generation of Protein Conformational Ensembles with Backbone-to-Side-Chain Transformers

Excitement at the prospect of using data-driven generative models to sample configurational ensembles of biomolecular systems stems from the extraordinary success of these models on a diverse set of high-dimensional sampling tasks. Unlike image generation or even the closely related problem of protein structure prediction, there are currently no data sources with sufficient breadth to parametrize generative models for conformational ensembles. To enable discovery, a fundamentally different approach to building generative models is required: models should be able to propose rare, albeit physical, conformations that may not arise in even the largest data sets. Here, in this work, we introduce a modular strategy to generate conformations based on “backmapping” from a fixed protein backbone that (1) maintains conformational diversity of the side chains and (2) couples the side-chain fluctuations using global information about the protein conformation. Our model combines simple statistical models of side-chain conformations based on rotamer libraries with the now ubiquitous transformer architecture to sample with atomistic accuracy. Together, these ingredients provide a strategy for rapid data acquisition and hence a crucial ingredient for scalable physical simulation with generative neural networks.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Co-scheduling Ensembles of In Situ Workflows

Molecular dynamics (MD) simulations are widely used to study large-scale molecular systems. HPC systems are ideal platforms to run these studies, however, reaching the necessary simulation timescale to detect rare processes is challenging, even with modern supercomputers. To overcome the timescale limitation, the simulation of a long MD trajectory is replaced by multiple short-range simulations that are executed simultaneously in an ensemble of simulations. Analyses are usually co-scheduled with these simulations to efficiently process large volumes of data generated by the simulations at runtime, thanks to in situ techniques. Executing a workflow ensemble of simulations and their in situ analyses requires efficient co- scheduling strategies and sophisticated management of computational resources so that they are not slowing down each other. In this paper, we propose an efficient method to co-schedule simulations and in situ analyses such that the makespan of the workflow ensemble is minimized. We present a novel approach to allocate resources for a workflow ensemble under resource constraints by using a theoretical framework modeling the workflow ensemble’s execution. We evaluate the proposed approach using an accurate simulator based on the WRENCH simulation framework on various workflow ensemble configurations. Results demonstrate the significance of co-scheduling simulations and in situ analyses that couple data together to benefit from data locality, in which inefficient scheduling decisions can lead to slowdown in makespan up to a factor of 30.

Do, Tu Mai Anh↗
NASA NTRS2015

Inconsistent Strategies to Spin up Models in CMIP5: Implications for Ocean Biogeochemical Model Performance Assessment

During the fifth phase of the Coupled Model Intercomparison Project (CMIP5) substantial efforts were made to systematically assess the skill of Earth system models. One goal was to check how realistically representative marine biogeochemical tracer distributions could be reproduced by models. In routine assessments model historical hindcasts were compared with available modern biogeochemical observations. However, these assessments considered neither how close modeled biogeochemical reservoirs were to equilibrium nor the sensitivity of model performance to initial conditions or to the spin-up protocols. Here, we explore how the large diversity in spin-up protocols used for marine biogeochemistry in CMIP5 Earth system models (ESMs) contributes to model-to-model differences in the simulated fields. We take advantage of a 500-year spin-up simulation of IPSL-CM5A-LR to quantify the influence of the spin-up protocol on model ability to reproduce relevant data fields. Amplification of biases in selected biogeochemical fields (O2, NO3, Alk-DIC) is assessed as a function of spin-up duration. We demonstrate that a relationship between spin-up duration and assessment metrics emerges from our model results and holds when confronted with a larger ensemble of CMIP5 models. This shows that drift has implications for performance assessment in addition to possibly aliasing estimates of climate change impact. Our study suggests that differences in spin-up protocols could explain a substantial part of model disparities, constituting a source of model-to- model uncertainty. This requires more attention in future model intercomparison exercises in order to provide quantitatively more correct ESM results on marine biogeochemistry and carbon cycle feedbacks.

histories↗
NASA NTRS2016

An Assessment of Multimodel Simulations for the Variability of Western North Pacific Tropical Cyclones and Its Association with ENSO

An assessment of simulations of the interannual variability of tropical cyclones (TCs) over the western North Pacific (WNP) and its association with El Niño-Southern Oscillation (ENSO), as well as a subsequent diagnosis for possible causes of model biases generated from simulated large-scale climate conditions, are documented in the paper. The model experiments are carried out by the Hurricane Work Group under the U.S. Climate Variability and Predictability Research Program (CLIVAR) using five global climate models (GCMs) with a total of 16 ensemble members forced by the observed sea surface temperature and spanning the 28-yr period from 1982 to 2009. The results show GISS and GFDL model ensemble means best simulate the interannual variability of TCs, and the multimodel ensemble mean (MME) follows. Also, the MME has the closest climate mean annual number of WNP TCs and the smallest root-mean-square error to the observation. Most GCMs can simulate the interannual variability of WNP TCs well, with stronger TC activities during two types of El Niño-namely, eastern Pacific (EP) and central Pacific (CP) El Niño-and weaker activity during La Niña. However, none of the models capture the differences in TC activity between EP and CP El Niño as are shown in observations. The inability of models to distinguish the differences in TC activities between the two types of El Niño events may be due to the bias of the models in response to the shift of tropical heating associated with CP El Niño.

climate models↗
NASA NTRS2014

Hybrid Data Assimilation without Ensemble Filtering

The Global Modeling and Assimilation Office is preparing to upgrade its three-dimensional variational system to a hybrid approach in which the ensemble is generated using a square-root ensemble Kalman filter (EnKF) and the variational problem is solved using the Grid-point Statistical Interpolation system. As in most EnKF applications, we found it necessary to employ a combination of multiplicative and additive inflations, to compensate for sampling and modeling errors, respectively and, to maintain the small-member ensemble solution close to the variational solution; we also found it necessary to re-center the members of the ensemble about the variational analysis. During tuning of the filter we have found re-centering and additive inflation to play a considerably larger role than expected, particularly in a dual-resolution context when the variational analysis is ran at larger resolution than the ensemble. This led us to consider a hybrid strategy in which the members of the ensemble are generated by simply converting the variational analysis to the resolution of the ensemble and applying additive inflation, thus bypassing the EnKF. Comparisons of this, so-called, filter-free hybrid procedure with an EnKF-based hybrid procedure and a control non-hybrid, traditional, scheme show both hybrid strategies to provide equally significant improvement over the control; more interestingly, the filter-free procedure was found to give qualitatively similar results to the EnKF-based procedure.

Kalman Filter↗
NASA NTRSDate not supplied

Multi-source Estimates of Land / Ocean Moisture Transport Variability over the Satellite Era

It is widely appreciated that atmospheric transport of water from the world’s oceans is a process key to planetary energy balance as well as Earth's habitability. What is not yet clear is the extent of variability in moisture transports, the relative importance of interdecadal variability versus climate change signals, and importantly, our ability to quantify these changes. This work assesses variations in moisture transport variability during the satellite era (~1980 to present) by comparing several different estimates. (i) The most direct estimate is the vertically integrated flux convergence of moisture from reanalyses which use observed wind and moisture information. (ii) One alternative estimate comes from P-ET over land taken from global hydrologic models constrained with precipitation and near-surface meteorology. Here we use an ensemble of six models. An adjunct to this method is to employ satellite derive ET (e.g., GLEAM or DOLCE). (iii) Complementary to this is E-P over the global oceans derived from satellite estimates of P such as TRMM, GPM and GPCP and SeaFlux V3 or J-OFURO3 estimates of E, all relying heavily upon microwave measurements. Transport between land and oceans must essentially balance at monthly scales, i.e., vanish globally. (iv) a fourth perspective comes from estimate of terrestrial RO + storage rate, delta S. G-RUN Ensemble which uses observed streamflow and P measurements to calibrate a statistical model provides the former while GRACE, GRACE-FO provide total water storage anomalies used to calculate storage rate changes. GRACE REC uses GRACE data to train a precipitation-driven statistical model to extend storage estimates before the GRACE era. (All of these alternatives to reanalysis estimates also consider the small atmospheric column water vapor contribution.) We examine the transport changes from these three different methodologies, their relative accuracies and discuss the origin of their differences. Regional trends in moisture flux divergence and their role in multi-decadal trends are considered. Interannual variability arising in connection with ENSO variability is a dominant signal, driven largely by P changes. Trends since 1980 include reductions in moisture delivery to the western U.S., eastern Brazil, and central Africa with recovery of moisture convergence to the Sahel and parts of eastern North America.

Franklin Robertson↗
DOE OSTI2021

PreMevE Update: Forecasting Ultra-Relativistic Electrons Inside Earth's Outer Radiation Belt

Energetic electrons inside Earth's Van Allen belts pose a major radiation threat to spaceborne electronics that often play vital roles in modern society. Ultra-relativistic electrons with energies greater than or equal to two megaelectron-volt (MeV) are of particular interest, and thus forecasting these ≥2 MeV electrons has a significant meaning to all space sectors. Here, we update the latest development of the predictive model for MeV electrons in the outer radiation belt. The new version, called PREdictive MEV Electron (PreMevE)-2E, forecasts ultra-relativistic electron flux distributions across the outer belt, with no need for in situ measurements of the trapped MeV electron population except at the geosynchronous orbit (GEO). Model inputs include precipitating electrons observed in low-Earthorbits by NOAA satellites, upstream solar wind speeds and densities from solar wind monitors, as well as ultra-relativistic electrons measured by one Los Alamos GEO satellite. We evaluated 32 supervised machine learning models that fall into four different classes of linear and neural network architectures, and successfully tested ensemble forecasting by using groups of top-performing models. All models are individually trained, validated, and tested by in situ electron data from NASA's Van Allen Probes mission. It is shown that the final ensemble model outperforms individual models at most L-shells, and this PreMevE-2E model can provide 25-h (~1-day) and 50-h (~2-day) forecasts with high mean performance efficiency and correlation values. Our results also suggest that this new model is dominated by nonlinear components at L-shells <~4 for ultra-relativistic electrons, different from the dominance of linear components for 1 MeV electrons as previously discovered.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

GMRES with embedded ensemble propagation for the efficient solution of parametric linear systems in uncertainty quantification of computational models

In a previous work, embedded ensemble propagation was proposed to improve the efficiency of sampling-based uncertainty quantification methods of computational models on emerging computational architectures. It consists of simultaneously evaluating the model for a subset of samples together, instead of evaluating them individually. A first approach introduced to solve parametric linear systems with ensemble propagation is ensemble reduction. In Krylov methods for example, this reduction consists in coupling the samples together using an inner product that sums the sample contributions. Ensemble reduction has the advantages of being able to use optimized implementations of BLAS functions and having a stopping criterion which involves only one scalar. However, the reduction potentially decreases the rate of convergence due to the gathering of the spectra of the samples. In this paper, we investigate a second approach: ensemble propagation without ensemble reduction in the case of GMRES. This second approach solves each sample simultaneously but independently to improve the convergence compared to ensemble reduction. This raises two new issues which are solved in this paper: the fact that optimized implementations of BLAS functions cannot be used anymore and that ensemble divergence, whereby individual samples within an ensemble must follow different code execution paths, can occur. We tackle those issues by implementing a high-performing ensemble GEMV and by using masks. The proposed ensemble GEMV leads to a similar cost per GMRES iteration for both approaches, i.e. with and without reduction. For illustration, we study the performances of the new linear solver in the context of a mesh tying problem. Furthermore, this example demonstrates improved ensemble propagation speed-up without reduction.

BLAS↗
DOE OSTI2024

Evaluating Ensemble Predictions of South Asian Monsoon Low Pressure System Genesis

Abstract Synoptic-scale vortices known as monsoon low pressure systems (LPSs) frequently produce intense precipitation and hydrological disasters in South Asia, so accurately forecasting LPS genesis is crucial for improving disaster preparedness and response. However, the accuracy of LPS genesis forecasts by numerical weather prediction models has remained unknown. Here, we evaluate the performance of two global ensemble models—the U.S. Global Ensemble Forecast System (GEFS) and the Ensemble Prediction System of the European Centre for Medium-Range Weather Forecasts (ECMWF)—in predicting LPS genesis during the years 2021–22. The GEFS successfully predicted about half the observed LPS genesis events 1–2 days in advance; the ECMWF model captured an additional 10% of observed genesis events. Both models had a false alarm ratio (FAR) of around 50% for 1–2-day lead times. In both ensembles, the control run typically exhibited a higher probability of detection (POD) of observed events and a lower FAR compared to the perturbed ensemble members. However, a consensus forecast, in which genesis is predicted when at least 20% of ensemble members forecast LPS formation, had POD values surpassing those of the control run for all lead times. Moreover, probabilistic predictions of genesis over the Bay of Bengal, where most LPSs form, were skillful, with the fraction of ensemble members predicting LPS formation over a 5-day lead time approximating the observed frequency of genesis, without any adjustment or bias correction.

Suhas, D. L.↗