Search NASA⌕ Search

SEARCH · Search NASA

Results for “simulation datasets”

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 973 records · Page 54

Myna: Connecting powder bed fusion build data to simulation tools for digital twin applications

Additive manufacturing (AM), as a digital process, can generate a detailed digital thread linking a part’s design and manufacturing to its operational performance. As AM systems advance, an increasing amount of process data is stored in manufacturing databases. In principle, this data can be utilized by simulation-based digital twin approaches, such as real-time process control and asynchronous post-processing guidance. However, few tools currently exist for systematically integrating digital thread data with computational tools. Here, in this study, we propose a software package, called Myna, for connecting data from powder bed fusion processes to simulation tools. The utility of such a platform is demonstrated using build data from the Oak Ridge National Laboratory Manufacturing Demonstration Facility “Peregrine v2023-10” public dataset to automatically configure and run 54 semi-analytical 3DThesis melt pool simulations, 78 numerical Additive FOAM melt pool simulations, and 3 ExaCA microstructure simulations. The simulated, spatially registered microstructures are then compared directly with electron backscatter diffraction characterization of the corresponding as-built part locations. The resulting simulated microstructure showed variation as a function of process parameters, particularly stripe width; however, the experimental data had little variation between the microstructure texture and grain size resulting from different processing conditions. Analysis of the discrepancies suggest that it is possible a two-phase ferritic-austenitic solidification model is needed to accurately predict grain size and texture for certain stainless steel 316L feedstock compositions under powder bed fusion conditions, providing direction for future research. As illustrated here, due to the number and complexity of the simulations involved in AM process-structure–property predictions, automated methods to connect process data and simulations will remain necessary tools for testing hypotheses and implementing digital twin applications.

Knapp, Gerald L. [Oak Ridge National Laboratory (O↗

Experimental Design and Laboratory Characterization of a Medium- and High-Efficiency Rooftop Unit for use in Building Energy Simulations

Commercial rooftop units that incorporate variable speed components are quickly entering into the market place; however, the design and modeling tools that calculate the energy benefits of these units cannot properly estimate performance. This is due to two factors: 1) There are insufficient measured performance data of real systems that can be entered into these simulations, and 2) building energy simulators are unable to properly interpret performance data. Therefore, publicly available datasets designed to encompass the entire performance map coupled with control sequence information are needed as inputs for these modelling programs to accurately estimate energy savings and promote variable speed RTUs. As part of the Verification and Uncertainty Characterization for Energy Simulation milestone funded by the Department of Energy, the National Renewable Energy Laboratory recorded and published data across the entire performance range for two commercially available RTUs; one with a two-stage compressor and variable speed supply fan, and another with a fully variable speed compressor, supply fan and condenser fan. This paper details 1) how the design of experiments was created as well as how the units were controlled, and 2) the operational limitations of the two units. The paper will also discuss methods for building energy simulation programs to properly interpret these data sets and integrate them into building energy simulation models.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Upgrade and benchmarking of a 4D treatment planning system for scanned ion beam therapy

Purpose: Upgrade and benchmarking of a research 4D treatment planning system (4DTPS) suitable for realistic patient treatment planning and treatment simulations taking into account specific requirements for scanned ion beam therapy, i.e., modeling of dose heterogeneities due to interplay effects and range changes caused by patient motion and dynamic beam delivery. Methods: The 4DTPS integrates data interfaces to 4D computed tomography (4DCT), deformable image registration and clinically used motion monitoring devices. The authors implemented a novel data model for 4D image segmentation using Boolean mask volume datasets and developed an algorithm propagating a manually contoured reference contour dataset to all 4DCT phases. They further included detailed treatment simulation and dose reconstruction functionality, based on the irregular patient motion and the temporal structure of the beam delivery. The treatment simulation functionality was validated against experimental data from irradiation of moving radiographic films in air, 3D moving ionization chambers in a water phantom, and moving cells in a biological phantom with a scanned carbon ion beam. The performance of the program was compared to results obtained with predecessor programs. Results: The measured optical density distributions of the radiographic films were reproduced by the simulations to (-2 {+-} 12)%. Compared to earlier versions of the 4DTPS, the mean agreement improved by 2%, standard deviations were reduced by 7%. The simulated dose to the moving ionization chambers in water showed an agreement with the measured dose of (-1 {+-} 4)% for the typical beam configuration. The mean deviation of the simulated from the measured biologically effective dose determined via cell survival was (617 {+-} 538) mGy relative biological effectiveness corresponding to (10 {+-} 9)%. Conclusions: The authors developed a research 4DTPS suitable for realistic treatment planning on patient data and capable of simulating dose delivery to a moving patient geometry for scanned ion beams. The accuracy and reliability of treatment simulations improved considerably with respect to earlier versions of the 4DTPS.

60 APPLIED LIFE SCIENCES↗

Data for "Divergent urban land trajectories under alternative population projections within the shared socioeconomic pathways"

Datasets supporting findings and visualizations behind McManamay et al. (2024) Divergent urban land trajectories under alternative population projections within the shared socioeconomic pathways. Environmental Research Letters, DOI: 10.1088/1748-9326/ad2eec Using a spatial modeling experiment at high resolution (1-km), this study compared how two alternative US population projections, varying in the spatially explicit nature of demographic patterns and migration, affect urban land dynamics simulated by the Spatially Explicit, Long-term, Empirical City development (SELECT) model for SSP2, SSP3, and SSP5. The numerical experiment, inputs and outputs are fully described in McManamay et al. (2024). The datasets summarize SELECT model simulations within urban areas and rural areas of the conterminous United States. For code to reproduce the results, please refer to https://github.com/IMMM-SFA/mcmanamay_etal_2024_erl Please refer to the README file provided in Files for more details. Descriptions of the datasets are provided below. Dataset(s) Descriptions: Urban_Land_delta_UA_County.csv: Urban land area for original (default) population, updated population, and urban land delta (difference between the two) according to Shared Socioeconomic Pathways (SSPs) and year. Urban land is summarized for U.S. counties (FIPS code) and urban areas (GEOID) as unique areas within various counties. JO_Gao_pop_ufdelta.csv: Comparison of population differences and urban land differences between the origial (default) and updated population projections. City_Case_studies.csv: Differences in urban land areas arising from different population projections summarized for selected cities (Atlanta, Los Angeles, Houston, and New York) and their surrounding rural areas UA_county_FID.zip: Zipped folder of .tiff file depicting unique combination of urban areas and US counties used for summarizing urban land differences in urban and rural areas based on different population projections. ULD_clusters.zip: Zipped folder of .csv files depicting normalized changes in urban land delta (difference in urban land arising from population projections) and the respective Ward's hierarchical clusters, which group urban and rural areas based on similarities in temporal trends. Urban_Percent_change_2100.zip: Zipped folder of .csv files depicting percent changes in urban land area for year 2100 based on differences in the original (default) population and updated population. Each file is for SSPs and urban or rural areas.

Land↗

2021 Smoky Mountains Conference Data Challenge Synthetic-to-Real Domain Adaptation for Autonomous Driving Dataset

The dataset is comprised of both real and synthetic images from a vehicle's forward-facing camera. Each camera image is accompanied by a corresponding pixel-level semantic segmentation image (all files are .png files). In total, the dataset contains 5600 images in the training/validation set and 1400 images in the testing set. The training dataset contains mostly synthetic RGB images collected with a wide range of weather and lighting conditions using the CARLA simulator [1]. In addition, the training data also includes a small pre-selected subset of data from the Cityscapes training dataset – which is comprised of RGB-segmentation image pairs from driving scenarios in various European cities [2]. The testing data is split into three sets. The first set contains synthetic CARLA images with weather/lighting conditions that were not present in the training set. The second set is a subset of the Cityscapes testing dataset. Finally, the third set is an unknown testing set which will not be revealed to the participants until after the submission deadline. [1] Dosovitskiy, A., Ros, G., Codevilla, F., Lopez, A., and Koltun, V. (2017, October). CARLA: An open urban driving simulator. In Conference on robot learning (pp. 1-16). PMLR. [2] Cordts, M., Omran, M., Ramos, S., Rehfeld, T., Enzweiler, M., Benenson, R., ... and Schiele, B. (2016). The cityscapes dataset for semantic urban scene understanding. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3213-3223).

99 GENERAL AND MISCELLANEOUS↗

High Energy Physics Exascale Requirements Review. An Office of Science review sponsored jointly by Advanced Scientific Computing Research and High Energy Physics, June 10-12, 2015, Bethesda, Maryland

The U.S. Department of Energy (DOE) Office of Science (SC) Offices of High Energy Physics (HEP) and Advanced Scientific Computing Research (ASCR) convened a programmatic Exascale Requirements Review on June 10–12, 2015, in Bethesda, Maryland. This report summarizes the findings, results, and recommendations derived from that meeting. The high-level findings and observations are as follows. Larger, more capable computing and data facilities are needed to support HEP science goals in all three frontiers: Energy, Intensity, and Cosmic. The expected scale of the demand at the 2025 timescale is at least two orders of magnitude — and in some cases greater — than that available currently. The growth rate of data produced by simulations is overwhelming the current ability of both facilities and researchers to store and analyze it. Additional resources and new techniques for data analysis are urgently needed. Data rates and volumes from experimental facilities are also straining the current HEP infrastructure in its ability to store and analyze large and complex data volumes. Appropriately configured leadership-class facilities can play a transformational role in enabling scientific discovery from these datasets. A close integration of high-performance computing (HPC) simulation and data analysis will greatly aid in interpreting the results of HEP experiments. Such an integration will minimize data movement and facilitate interdependent workflows. Long-range planning between HEP and ASCR will be required to meet HEP’s research needs. To best use ASCR HPC resources, the experimental HEP program needs (1) an established, long-term plan for access to ASCR computational and data resources, (2) the ability to map workflows to HPC resources, (3) the ability for ASCR facilities to accommodate workflows run by collaborations potentially comprising thousands of individual members, (4) to transition codes to the next-generation HPC platforms that will be available at ASCR facilities, (5) to build up and train a workforce capable of developing and using simulations and analysis to support HEP scientific research on next-generation systems.

97 MATHEMATICS AND COMPUTING↗

Representation of Extratropical Cyclones, Blocking Anticyclones, and Alpine Circulation Types in Multiple Reanalyses and Model Simulations

Atmospheric circulation types, blockings, and cyclones are prime features of the extratropical flow and key to understanding the climate system. This study intercompares the representation of these features in 10 reanalyses and in an ensemble of 30 climate model simulations between 1980 and 2005. Both modern, full-input reanalyses and century-long, surface-input reanalyses are examined. Modern full-input reanalyses agree well on key statistics of blockings, cyclones, and circulation types. Yet, the intensity and depth of cyclones vary among them. Reanalyses with higher horizontal resolution show higher cyclone center densities and more intense cyclones. For blockings, no strict relationship is found between frequency or intensity and horizontal resolution. Full-input reanalyses contain more intense blocking, compared to surface-input reanalyses. Circulation-type classifications over central Europe show that both versions of the Twentieth Century Reanalysis dataset contain more easterlies and fewer westerlies than any other reanalysis, owing to their high pressure bias over northeast Europe. The temporal correlation of annual circulation types over central Europe and blocking frequencies over the North Atlantic–European domain between reanalyses is high (around 0.8). The ensemble simulations capture the main characteristics of midlatitudinal atmospheric circulation. Circulation types of westerlies to northerlies over central Europe are overrepresented. There are too few blockings in the higher latitudes and an excess of cyclones in the midlatitudes. Other characteristics, such as blocking amplitude and cyclone intensity, are realistically represented, making the ensemble simulations a rich dataset to assess changes in climate variability.

54 ENVIRONMENTAL SCIENCES↗

SU-E-T-26: A Dosimetric Comparison of Two Treatment Setups for Lung Stereotactic Body Radiation Therapy (SBRT) Patients

Purpose: To assess the feasibility of treating lung SBRT patients with the ipsilateral arm adducted beside the body instead of elevated above the head. Methods: Patients receiving lung SBRT at our institution are typically treated with both arms raised above their head. However, several patients had difficulty maintaining their arms in an elevated position. In this study, lung SBRT patients who underwent PET-CT imaging with an adducted ipsilateral arm were selected to investigate the dosimetric effects of this treatment setup. PET-CT datasets were fused with treatment planning CT images to simulate the adducted arm position. One VMAT treatment plan was created per patient using the Pinnacle treatment planning system. Plans were optimized to achieve minimal dose to the ipsilateral arm while keeping the target coverage and critical structure doses within clinical limits. The target dose coverage, conformity index (CI) for the target, and DVHs of critical structures for the adducted arm plan were calculated. These parameters were compared with the clinical plan and reported along with the mean and maximum doses of the ipsilateral arm. Results: The target coverage, CI and DVHs for the adducted arm plans of two patients (one with peripheral lesion and one with central lesion) were comparable with the clinical plans. Dose constraints for the chest wall limited further reduction of ipsilateral arm doses for the peripheral lesion plan. The mean ipsilateral arm doses for the central and peripheral lesions were 0.33 Gy and 2.4 Gy, respectively. The maximum ipsilateral arm doses for the central and peripheral lesions were 1.0 Gy and 6.2 Gy, respectively. Conclusion: The results suggested patients with central lung SBRT tumors were more suitable for treatment with the adducted arm position. More patients with various lung tumor locations will be studied to find optimal tumor locations for treatment with this arm position.

60 APPLIED LIFE SCIENCES↗

Simulating the Trajectory and Biomass Growth of Free-Floating Macroalgal Cultivation Platforms along the U.S. West Coast

Trajectory tracking and macroalgal growth models were coupled to support a novel macroalgae-harvesting concept known as the Nautical Off-shore Macroalgal Autonomous Device (NOMAD). The NOMAD consists of 5 km long carbon-fiber longlines that are seeded and free float southward along the U.S. West Coast for approximately 3 months before harvesting off the California coast, taking advantage of favorable environmental conditions. The trajectory and macroalgal growth models were applied to answer planning questions pertinent to the techno-economic analysis such as identifying the preferred release location, approximate pathway, timing until harvest, and estimated growth. Trajectories were determined with the General NOAA Operational Modeling Environment (GNOME) model, using 11 years of current and wind data, determining probabilities by running nearly 40,000 Monte Carlo simulations varying the start time and location. An accompanying macroalgal growth model was used to estimate the growth of macroalgae based on the trajectory tracks and environmental forcing products, including light, temperature and nutrients. Model results show that NOMAD lines transit south in the months of April to September due to seasonal currents, taking approximately 3 months to reach Southern California. During transit, NOMAD lines are dispersed but typically avoid beaching or passing through marine sanctuaries. NOMAD lines can yield up to 30 kg wet weight per meter of cultivation line.

09 BIOMASS FUELS↗

Foundation models for atomistic simulation of chemistry and materials

Conventional computational methods for modeling chemical and materials systems are limited by system size and timescale, forcing a trade-off between quantum-mechanical accuracy and the sampling needed for realistic observables. Large language and vision foundation models — pre-trained on massive datasets using transformer architectures — have revolutionized many fields. It is thus interesting to ask whether a foundation model — subject to suitable data, parameter scaling and training — could enable learned simulations of chemistry and materials. Here, in this study, we review the field of machine-learned interatomic potentials (MLIPs) and posit that scaling up large and diverse chemical and materials datasets and highly expressive architectures using advanced training strategies should result in models that are: more efficient, transferable, robust to out-of-distribution scenarios, and easier to fine-tune to a variety of downstream physical observables than models trained from scratch on small datasets corresponding to specific, targeted atomistic simulation tasks. We provide specific criteria for creating such large-scale MLIP foundation models, coordinated strategies for their development, evaluation and deployment, and highlight potential emergent capabilities that could transform predictive simulations in chemistry and materials science and accelerate discovery across multiple technological domains.

Yuan, Eric C.-Y. [University of California, Berkel↗

Fast physics-based launcher optimization for electron cyclotron current drive

With the increased urgency to design fusion pilot plants, fast optimization of electron cyclotron current drive (ECCD) launchers is paramount. Traditionally, this is done by coarsely sampling the 4D parameter space of possible launch conditions consisting of (1) the launch location (constrained to lie along the reactor vessel), (2) the launch frequency, (3) the toroidal launch angle, and (4) the poloidal launch angle. For each initial condition, a ray-tracing simulation is performed to evaluate the ECCD efficiency. Unfortunately, this approach often requires a large number of simulations (sometimes millions in extreme cases) to build up a dataset that adequately covers the plasma volume, which must then be repeated every time the design point changes. Here we adopt a different approach. Rather than launching rays from the plasma periphery and hoping for the best, we instead directly reconstruct the optimal ray for driving current at a given flux surface using a reduced physics model coupled with a commercial ray-tracing code. Repeating this throughout the plasma volume requires only hundreds of simulations, constituting a significant speedup. The new method is validated on two separate example tokamak profiles, and is shown to reliably drive localized current at the specified flux surface with the same optimal efficiency as obtained from the traditional approach.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Characterizing How Meteorological Forcing Selection and Parameter Uncertainty Influence Community Land Model Version 5 Hydrological Applications in the United States

Despite the increasing use of large-scale Land Surface Models (LSMs) in predicting hydrological responses in extreme conditions, there's a critical gap in understanding the uncertainties in these predictions. This study addresses this gap through a detailed diagnostic evaluation of the uncertainties arising from meteorological forcing selection and model parametrization in hydrological simulations of the Community Land Model version 5 (CLM5). CLM5 is configured at a spatial scale of about 12-km to simulate runoff processes for 464 headwater watersheds, selected from the Catchment Attributes for Large-Sample Studies (CAMELS) dataset to be representative of physiographic and climatic gradients across the conterminous United States. For each watershed, CLM5 is driven by five commonly used gridded forcing datasets in combination with a large ensemble (> 1200) of key CLM5 hydrologic parameters. Our results suggest that uncertainty in CLM5 runoff simulations resulting from both forcing and parametric sources is markedly higher in arid regions, e.g., Great Plains and Midwest regions. Uncertainty in low flow is dominated by parametric uncertainty, while the selection of meteorological forcing contributes more dominantly to high flow and seasonal flows during fall and spring. Our analysis also demonstrates that the selection of forcing datasets and the metrics used to calibrate CLM5 significantly impact the model’s predictive accuracy in extreme event severity for both floods and droughts. Overall, the results from this study highlight the need to understand and account for forcing and parametric uncertainties in CLM5 simulations, particularly for hazard and risk assessments addressing hydrologic extremes.

54 ENVIRONMENTAL SCIENCES↗

Quantification of Arctic Soil and Permafrost Properties Using Ground-Penetrating Radar and Electrical Resistivity Tomography Datasets

Improving understanding of Arctic ecosystem climate feedback and parameterization of models that simulate freeze-Thaw dynamics require advances in quantifying soil and snow properties. Due to the significant spatiotemporal variability of soil properties and the limited information provided by point-scale measurements (e.g., cores), geophysical methods hold potential for improving soil and permafrost characterization. In this study, we evaluate the use of a ground-penetrating radar (GPR) to estimate thaw layer thickness, snow depth, and ice-wedge characteristics in an ice-wedge-dominated tundra region near Barrow, AK, USA. To this end, we analyze GPR and point-scale measurements collected along several parallel transects at the end of the growing season and the end of frozen season. In addition, we compare the structural information extracted from the GPR data with electrical resistivity tomography (ERT) information about ice-wedge characteristics. Our study generally highlights the value of GPR data collected in the frozen season, when conditions lead to the improved GPR signal-To-noise ratio, facilitate data acquisition, and reduce acquisition-related ecosystem disturbance relative to growing season. We document for the first time that GPR data collected during the frozen season can provide reliable estimates of active layer thickness and geometry of ice wedges. We find that the ice-wedge geometry extracted from GPR data collected during the frozen season is consistent with ERT data, and that GPR data can be used to constrain the ERT inversion. Consistent with recent studies, we also find that GPR data collected during the frozen season can provide good estimates of snow thickness, and that GPR data collected during the growing season can provide reliable estimate thaw depth. As a result, our quantification of the value of the GPR and ERT data collected during growing and frozen seasons paves the way for coupled inversion of the datasets to improve understanding of permafrost variability.

54 ENVIRONMENTAL SCIENCES↗

Subordinated Gaussian Processes for Solar Irradiance

Traditionally the power grid has been a one-way street with power flowing from large transmission-connected generators through the distribution network to consumers. This paradigm is changing with the introduction of distributed renewable energy resources (DERs), and with it, the way the grid is managed. There is currently a dearth of high fidelity solar irradiance datasets available to help grid researchers understand how expansion of DERs could affect future power system operations. Realistic simulations of by-the-second solar irradiances are needed to study how DER variability affects the grid. Irradiance data are highly non-stationary and non-Gaussian, and even modern time series models are challenged by their distributional properties. We develop a subordinated non-Gaussian stochastic model whose simulations realistically capture the distribution and dependence structure in measured irradiance. We illustrate our approach on a fine resolution dataset from Hawaii, where our approach outperforms standard nonlinear time series models.

MATHEMATICS AND COMPUTING,SOLAR ENERGY↗

Canopus: A Paradigm Shift Towards Elastic Extreme-Scale Data Analytics on HPC

Scientific simulations on high performance computing (HPC) platforms generate large quantities of data. To bridge the widening gap between compute and I/O, and enable data to be more efficiently stored and analyzed, simulation outputs need to be refactored, reduced, and appropriately mapped to storage tiers. However, a systematic solution to support these steps has been lacking on the current HPC software ecosystem. To that end, this paper develops Canopus, a progressive JPEGlike data management scheme for storing and analyzing big scientific data. It co-designs the data decimation, compression and data storage, taking the hardware characteristics of each storage tier into considerations. With reasonably low overhead, our approach refactors simulation data into a much smaller, reduced-accuracy base dataset, and a series of deltas that is used to augment the accuracy if needed. The base dataset and deltas are compressed and written to multiple storage tiers. Data saved on different tiers can then be selectively retrieved to restore the level of accuracy that satisfies data analytics. Thus, Canopus provides a paradigm shift towards elastic data analytics and enables end users to make trade-offs between analysis speed and accuracy on-the-fly. We evaluate the impact of Canopus on unstructured triangular meshes, a pervasive data model used by scientific modeling and simulations. In particular, we demonstrate the progressive data exploration of Canopus using the “blob detection” use case on the fusion simulation data.

Lu, Tao↗

Global patterns of extreme drought-induced loss in land primary production: Identifying ecological extremes from rain-use efficiency

Quantifying the ecological patterns of loss of ecosystem function in extreme drought is important to understand the carbon exchange between the land and atmosphere. Rain-use efficiency [RUE; gross primary production (GPP)/precipitation] acts as a typical indicator of ecosystem function. Here in this study, a novel method based on maximum rain-use efficiency (RUE max ) was developed to detect losses of ecosystem function globally. Three global GPP datasets from the MODIS remote sensing data (MOD17), ground upscaling FLUXNET observations (MPI-BGC), and process-based model simulations (BESS), and a global gridded precipitation product (CRU) were used to develop annual global RUE datasets for 2001–2011. Large, well-known extreme drought events were detected, e.g. 2003 drought in Europe, 2002 and 2011 drought in the U.S., and 2010 drought in Russia. Our results show that extreme drought-induced loss of ecosystem function could impact 0.9% ± 0.1% of earth's vegetated land per year and was mainly distributed in semi-arid regions. The reduced carbon uptake caused by functional loss (0.14 ± 0.03 PgC/yr) could explain >70% of the interannual variation in GPP in drought-affected areas (p ≤ 0.001). Our results highlight the impact of ecosystem function loss in semi-arid regions with increasing precipitation variability and dry land expansion expected in the future.

54 ENVIRONMENTAL SCIENCES↗

Aridity-Dependent Land Surface Skin Temperature Biases in CMIP5/6

Land surface skin temperature, a critical indicator of climate change, connects the water and energy cycles between the land and the atmosphere. Here, in this study, we evaluate the simulations of land surface skin temperature from the Coupled Model Intercomparison Project Phase 5 (CMIP5) and CMIP6 models with satellite-based datasets and reanalysis. We find systematic cold skin temperature biases over arid regions in CMIP5/CMIP6 simulations. Over arid and semi-arid regions, latent heat biases drive skin temperature biases by evaporative cooling. Over humid regions, surface downward shortwave and albedo biases are relatively more critical. Spatial patterns of biases remain similar in the latest CMIP6 simulations, suggesting systematic biases in land-atmosphere interactions. These biases need to be corrected or considered while using models for future projections.

54 ENVIRONMENTAL SCIENCES↗

ClimSim: A large multi-scale dataset for hybrid physics-ML climate emulation

Modern climate projections lack adequate spatial and temporal resolution due to computational constraints. A consequence is inaccurate and imprecise predictions of critical processes such as storms. Hybrid methods that combine physics with machine learning (ML) have introduced a new generation of higher fidelity climate simulators that can sidestep Moore’s Law by outsourcing compute-hungry, short, high-resolution simulations to ML emulators. However, this hybrid ML-physics simulation approach requires domain-specific treatment and has been inaccessible to MLexperts because of lack of training data and relevant, easy-to-use workflows. Wepresent ClimSim, the largest-ever dataset designed for hybrid ML-physics research. It comprises multi-scale climate simulations, developed by a consortium of climate scientists and ML researchers. It consists of 5.7 billion pairs of multivariate input and output vectors that isolate the influence of locally-nested, high-resolution, high-fidelity physics on a host climate simulator’s macro-scale physical state. The dataset is global in coverage, spans multiple years at high sampling frequency, and is designed such that resulting emulators are compatible with downstream coupling into operational climate simulators. We implement a range of deterministic and stochastic regression baselines to highlight the ML challenges and their scoring. The data (https://huggingface.co/datasets/LEAP/ClimSim_high-res2) and code(https://leap-stc.github.io/ClimSim)arereleasedopenlytosupport the development of hybrid ML-physics and high-fidelity climate simulations for the benefit of science and society.

artificial intelligence, machine learning↗