Search NASA⌕ Search

SEARCH · Search NASA

Results for “Small scale processes”

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 631 records · Page 35

Advanced Characterization of Fuel Pellet and Cladding from the HB Robinson NPP

In order to predict the behavior and response of materials to extreme environments of a nuclear reactor, comprehensive models have been developed over the past several decades. These models are typically validated against carefully designed single or double variable experiments which may be systematically altered in order build up a model of behavior under the conditions being examined. In reality, this does not necessarily compare well with the real application situations whereby the behavior of the material under irradiation is determined by multiple variables changing simultaneously e.g. temperature, irradiation flux and fluence, and pressure. This is particularly so when considering the highly dynamic environment of a fuel assembly within a nuclear power plant (NPP) core. This work aims to provide near-atomic resolution microstructural and chemical data of a fuel/clad specimen that has been used in service within a commercial NPP. Atom probe tomography and S/TEM characterization of the cladding showed the presence of small scale clusters containing Fe and Cr, along with segregation of these minor solutes to dislocations. SEM examination of the fuel pellet showed relatively large fission gas bubbles that had burst due to the metallographic polishing process to prepare the sample for examination. SEM-EDX mapping revealed the presence of Xr bubbles located beneath the surface that still retained their Xe concentration. Metallic precipitates were also observed. S/TEM-EDX of the fuel material also revealed the presence of small metallic precipitates. In addition, small Xe bubbles were also found to be present in the material, lying along grain boundaries only. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Discrete Event Modeling and Massively Parallel Execution of Epidemic Outbreak Phenomena

In complex phenomena such as epidemiological outbreaks, the intensity of inherent feedback effects and the significant role of transients in the dynamics make simulation the only effective method for proactive, reactive or post-facto analysis. The spatial scale, runtime speed, and behavioral detail needed in detailed simulations of epidemic outbreaks make it necessary to use large-scale parallel processing. Here, an optimistic parallel execution of a new discrete event formulation of a reaction-diffusion simulation model of epidemic propagation is presented to facilitate in dramatically increasing the fidelity and speed by which epidemiological simulations can be performed. Rollback support needed during optimistic parallel execution is achieved by combining reverse computation with a small amount of incremental state saving. Parallel speedup of over 5,500 and other runtime performance metrics of the system are observed with weak-scaling execution on a small (8,192-core) Blue Gene / P system, while scalability with a weak-scaling speedup of over 10,000 is demonstrated on 65,536 cores of a large Cray XT5 system. Scenarios representing large population sizes exceeding several hundreds of millions of individuals in the largest cases are successfully exercised to verify model scalability.

99 GENERAL AND MISCELLANEOUS↗

Fundamental catalytic challenges to design improved biomass conversion technologies

During the past ten years, there has been significant interest and investment in the study of catalytic conversion of biomass-derived feedstocks into renewable fuels and chemicals. In the United States, an estimated $25 billion has been spent by venture capitalists, industry, and government agencies during this period of time to commercialize “renewable technologies” including solar energy, wind power, batteries, and biofuels (e.g., cellulosic ethanol). Four societal factors are driving these investments including: (1) the increased price of crude oil; (2) concerns about global warming; (3) the desire to improve rural economies where biomass is produced, and; (4) national goals to become energy self-sufficient. Furthermore, underlying these efforts is the realization that lignocellulosic biomass is the only realistic, near-term and non-food-competitive source of renewable organic carbon. To this end, legislative efforts, such as the US Renewable Fuel Standards, have been implemented to create subsidies, tax credits, mandates, and loan guarantees to help bring renewable fuel technologies to market. However, the representative body of industrial efforts to this end has faced significant challenges, with several startup companies having commercialized biomass conversion technologies but having struggled to reach commercial scale; in fact, many of these companies have filed for bankruptcy. This situation is likely due to the challenges associated with scaling up unproven pioneer processes, as well as neophyte investors not understanding the decade-long time frames and the sheer amount of funding that is often required to bring chemical process technologies to market. Nevertheless, several emerging catalytic technologies have either entered the market place, or are currently demonstrating their technologies in fully integrated pilot plants. Commercial and near commercial technologies for second generation biomass conversion technologies to-date include, among others: biomass-derived jet and diesel fuel from both waste vegetable oils and ethanol; small scale production of renewable jet and diesel from landfill gases via Fischer-Tropsch synthesis; catalytic conversion of carbohydrates into gasoline and aromatics; hydropyrolysis of biomass into gasoline and diesel, and; catalytic conversion of wood into aromatics. To be successful, biomass conversion technologies must ultimately be able to compete economically with petroleum technologies, which are already operating at large commercial scales and have been practiced for decades. To this end, various factors must be considered when evaluating the potential for new processes to compete with incumbent technologies; factors such as regional variations in feedstock quality and availability, government policy, subsidies and tax rates, and the proprietary positions of the ancillary technologies that might support the process. By any measure, however, a critical metric of the economic potential of a process is its efficiency with respect to the yield of products from raw materials, and this metric of performance is directly related to atom efficiency of the underlying chemical reactions. Therefore, while promising biomass conversion technologies continue to demonstrate progress towards commercialization, there remains an important need for the catalysis community to aid in this effort by: (1) designing more active, selective and stable catalysts, and (2) elucidating a more detailed understanding of the catalytic chemistries underlying these processes. Indeed, the study of catalytic biomass conversion has grown tremendously during the past decade; and new or emerging technologies that are in the laboratory stage have allowed for biomass to be converted into a wider variety of commodity chemicals and the full range of liquid fuels that are produced from petroleum. The objective of this perspective is to highlight some of the ongoing, fundamental challenges with respect to the catalytic conversion of biomass into renewable products, and to provide insight as to how the catalysis community can overcome several of these challenges. It is our belief that, as an international academic community of catalysis researchers, we can (and must) work together to address these challenges, and to drive progress toward the production of next-generation renewable chemicals and fuels from biomass.

09 BIOMASS FUELS↗

Scalable economic extracellular synthesis of CdS nanostructured particles by a non-pathogenic thermophile

We report microbially facilitated synthesis of cadmium sulfide (CdS) nanostructured particles (NP) using anaerobic, metal-reducing Thermoanaerobacter sp. The extracellular CdS crystallites were <10 nm in size with yields of ~3 g/L of growth medium/month with demonstrated reproducibility and scalability up to 24 L. During synthesis, Thermoanaerobacter cultures reduced thiosulfate and sulfite salts to H2S, which reacted with Cd2+ cations to produce thermodynamically favored NP in a single step at 65oC with catalytic nucleation on the cell surfaces. Photoluminescence (PL) analysis of dry CdS NP revealed an exciton-dominated PL peak at 440 nm, having a narrow full width at half maximum of 10 nm. A PL spectrum of CdS NP produced by dissimilatory sulfur reducing bacteria was dominated by features associated with radiative exciton relaxation at the surface. High reproducibility of CdS NP PL features important for scale-up conditions was confirmed from test tubes to 24L batches at a small fraction of the manufacturing cost associated with conventional inorganic NP production processes.

Moon, Ji Won↗

Bringing weak transitions to light

Abstract Weak transitions between quantum states are of fundamental importance for a broad range of phenomena from analytical biochemistry to precision physics, but generally challenge experimental detection. Due to their small cross sections scaling with the absolute square of their transition matrix elements, spectroscopic measurements often fail in particular in the presence of competing background processes. Here we introduce a general concept to break this scaling law and enhance the transition probability by exploiting a stronger laser-coupled pathway to the same excited state. We demonstrate the concept experimentally by attosecond transient absorption spectroscopy in helium atoms. The quasi-forbidden transitions from the ground state 1s 2 to the weakly coupled doubly excited 2p3dandsp 2,4− states are boosted by an order of magnitude. Enhancing single-photon-suppressed transitions can find widespread applicability, from spectral diagnostics of complex molecules in life and chemical sciences to precision spectroscopy of weak transitions in metastable atomic nuclei in the search for new physics.

Science & Technology - Other Topics↗

Linking Small-Scale Flight Manoeuvers and Density Profiles to the Vertical Movement of Insects in the Nocturnal Stable Boundary Layer

Huge numbers of insects migrate over considerable distances in the stably-stratified night-time atmosphere with great consequences for ecological processes, biodiversity, ecosystem services and pest management. We used a combination of meteorological radar and lidar instrumentation at a site in Oklahoma, USA, to take a new look at the general assistance migrants receive from both vertical and horizontal airstreams during their long-distance flights. Movement in the nocturnal boundary layer (NBL) presents very different challenges for migrants compared to those prevailing in the daytime convective boundary layer, but we found that Lagrangian stochastic modelling is effective at predicting flight manoeuvers in both cases. A key feature for insect transport in the NBL is the frequent formation of a thin layer of fast-moving air – the low-level jet. Modelling suggests that insects can react rapidly to counteract vertical air movements and this mechanism explains how migrants are retained in the jet for long periods (e.g. overnight, and perhaps for several hours early in the morning). This results in movements over much longer distances than are likely in convective conditions, and is particularly significant for the reintroduction of pests to northern regions where they are seasonally absent due to low winter temperatures.

54 ENVIRONMENTAL SCIENCES↗

Magnetic Prandtl number dependence of the kinetic-to-magnetic dissipation ratio

Using direct numerical simulations of three-dimensional hydromagnetic turbulence, either with helical or non-helical forcing, we show that the kinetic-to-magnetic energy dissipation ratio always increases with the magnetic Prandtl number, i.e., the ratio of kinematic viscosity to magnetic diffusivity. This dependence can always be approximated by a power law, but the exponent is not the same in all cases. For non-helical turbulence, the exponent is around 1/3, while for helical turbulence it is between 0.6 and 2/3. In the statistically steady state, the rate of energy conversion from kinetic into magnetic by the dynamo must be equal to the Joule dissipation rate. We emphasize that for both small-scale and large-scale dynamos, the efficiency of the energy conversion depends sensitively on the magnetic Prandtl number, and thus on the microphysical dissipation process. To understand this behavior, we also study shell models of turbulence and one-dimensional passive and active scalar models. We conclude that the magnetic Prandtl number dependence is qualitatively best reproduced in the one-dimensional model as a result of dissipation via localized Alfvén kinks.

79 ASTRONOMY AND ASTROPHYSICS↗

Axion monodromy inflation with multi-natural modulations

We study parameter space in the axion monodromy inflation corrected by dynamically generated terms involving with the axion. The potential has the linear term with multiple sinusoidal functions, which play a role in generating modulations. We show that this potential leads both to a large tensor-to-scalar ratio r{sub T} ∼ 0.16 and to a large negative running of spectral index α{sub s} ∼ - (0.02 -0.03), ameliorating the tension between the result of the PLANCK and that of the BICEP2. To realize these results, a small hierarchy among dynamical scales is required whereas the decay constants in sinusoidal functions remain sub-Planckian in this model. We discuss also reheating process after the inflation in a bottom-up approach.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Dynamics of core accretion

In this paper, we perform three-dimensional hydrodynamic simulations of gas flowing around a planetary core of mass M pl = 10M ⊕ embedded in a near Keplerian background flow, using a modified shearing box approximation. We assume an ideal gas behaviour following an equation of state with a fixed ratio of the specific heats, γ = 1.42, consistent with the conditions of a moderate-temperature background disc with solar composition. No radiative heating or cooling is included in the models. We employ a nested grid hydrodynamic code implementing the ‘Piecewise Parabolic Method’ with as many as six fixed nested grids, providing spatial resolution on the finest grid comparable to the present-day diameters of Neptune and Uranus. We find that a strongly dynamically active flow develops such that no static envelope can form. The activity is not sensitive to plausible variations in the rotation curve of the underlying disc. It is sensitive to the thermodynamic treatment of the gas, as modelled by prescribed equations of state (either ‘locally isothermal’ or ‘locally isentropic’) and the temperature of the background disc material. The activity is also sensitive to the shape and depth of the core's gravitational potential, through its mass and gravitational softening coefficient. Each of these factors influences the magnitude and character of hydrodynamic feedback of the small-scale flow on the background, and we conclude that accurate modelling of such feedback is critical to a complete understanding of the core accretion process. The varying flow pattern gives rise to large, irregular eruptions of matter from the region around the core which return matter to the background flow: mass in the envelope at one time may not be found in the envelope at any later time. No net mass accretion into the envelope is observed over the course of the simulation and none is expected, due to our neglect of cooling. Except in cases of very rapid cooling however, as defined by locally isothermal or isentropic treatments, any cooling that does affect the envelope material will have limited consequences for the dynamics, since the flow quickly carries cooled material out of the core's environment entirely. The angular momentum of material in the envelope, relative to the core, varies both in magnitude and in sign on time-scales of days to months near the core and on time-scales a few years at distances comparable to the Hill radius. The dynamical activity contrasts with the largely static behaviour typically assumed within the framework of the core accretion model for Jovian planet formation. We show that material entering the dynamically active environment may suffer intense heating and cooling events the durations of which are as short as a few hours to a few days. Shorter durations are not observable in our work due to the limits of our resolution. Peak temperatures in these events range from T ~ 1000 K to as high as T ~ 3–4000 K, with densities ρ ~ 10 -9 to 10 -8 gcm -3 . These time-scales, densities and temperatures span a range consistent with those required for chondrule formation in the nebular shock model. Finally, we therefore propose that dynamical activity in the Jovian planet formation environment could be responsible for the production of chondrules and other annealed silicates in the solar nebula.

79 ASTRONOMY AND ASTROPHYSICS↗

Engineering Innovations, Challenges, and Opportunities for Lignocellulosic Biorefineries: Leveraging Biobased Polymer Production

Alternative polymer feedstocks are highly desirable to address environmental, social, and security concerns associated with petrochemical-based materials. Lignocellulosic biomass (LCB) has emerged as one critical feedstock in this regard because it is an abundant and ubiquitous renewable resource. LCB can be deconstructed to generate valuable fuels, chemicals, and small molecules/oligomers that are amenable to modification and polymerization. However, the diversity of LCB complicates the evaluation of biorefinery concepts in areas including process scale-up, production outputs, plant economics, and life-cycle management. We discuss aspects of current LCB biorefinery research with a focus on the major process stages, including feedstock selection, fractionation/deconstruction, and characterization, along with product purification, functionalization, and polymerization to manufacture valuable macromolecular materials. We highlight opportunities to valorize underutilized and complex feedstocks, leverage advanced characterization techniques to predict and manage biorefinery outputs, and increase the fraction of biomass converted into valuable products.

59 BASIC BIOLOGICAL SCIENCES↗

In‐operando thermal transport characterization of moving particle bed heat exchanger

The project aimed to develop modulated photothermal radiometry (MPR) for conducting in-operando measurements of thermal transport in moving particle bed heat exchangers (MPBE) located at Sandia National Laboratory (SNL). The plan was to initially conduct lab-scale measurements on a 1-meter-long particle channel at UCSD, followed by in- operando measurements at SNL on a 20 kWth MPBE prototype built for this project, and if time permitting, potential final deployment on SNL’s 1 MWth G3P3 MPBE. The first two goals were successfully achieved but due to changes with commissioning timeline of the G3P3 MPBE, in-operando measurements on the G3P3 MPBE could not be conducted. The measurement campaigns have resulted in some promising results and the MPR tool has been handed over to trained SNL personnel for future use. MPR measurements on UCSD’s 1-meter-long particle channel showed a peculiar length dependence of the particle bed thermal conductivity and near-wall thermal resistance. The bed thermal conductivity increased, and the near-wall thermal resistance decreased, with increasing distance from the channel inlet, both eventually reaching a plateau. This was found to be due to hydrostatic pressure screening by the channel walls, known as the “Janssen effect”, and is the first experimental study on its role in granular heat transfer. A 20 kWth small-scale MPBE prototype, the first of its kind with 3 mm particle channels and 2 mm sCO2 plates, was designed and manufactured by Vacuum Process Engineering (VPE) for this project. This MPBE prototype was installed in SNL’s HEX test stand and the MPR tool was integrated into the test stand for in-operando measurements. Several modifications have been made to the MPR setup for facile installation and easy operation at SNL. Measurements with Carbo HSP40/70 particles were conducted with various particle and sCO2 temperatures and flow rates. Further, apart from regular MPBE operation, fault scenarios such as plate warping, particle stagnation, and choking were also detected and MPR showed excellent sensitivity to the detection of such conditions. Through the course of this project, MPR has evolved from a sensitive laboratory technique to a rugged industrial tool with various applications in high temperature measurements and diagnostics.

14 SOLAR ENERGY↗

STEAM-SIDE OXIDE SCALE EXFOLIATION BEHAVIOR IN SUPERHEATERS AND REHEATERS

Advances in materials for power plants include not only new materials with higher-temperature capabilities, but also the use of current materials at increasingly higher temperatures. This latter activity builds on extensive experience of the performance of the various alloys, and provides a basis for identifying changes in alloy behavior with increasing temperature as well as understanding the factors that ultimately determine the maximum use temperatures of the different alloy classes. This paper presents results from an effort to model the exfoliation processes of steam-side oxide scales in a manner that describes as accurately as possible the evolution of strains in oxides growing inside small-diameter tubes subjected to large thermal gradients and to thermal transients typical of normal steam boiler operation. One way of portraying the results of such calculations is by plotting the evolving strains in a given oxide scale on an Exfoliation Diagram (of the type pioneered by Manning et al. of the British Central Electricity Research Laboratory) to determine the earliest time at which the trajectory of these strains intersects a criterion for scale failure. Understanding of how such strain trajectories differ among different alloys and are affected by the major variables associated with boiler operation has the potential to suggest boiler operating strategies to manage scale exfoliation, as well as to highlight the mode of scale failure and the limitations of each alloy. Preliminary results are presented of the strain trajectories calculated for alloys T22, T91, and TP347 subjected to the conditions experienced by superheaters under assumed boiler operating scenarios. For all three alloys the earliest predicted scale failures were associated with the increased strains developed during a boiler shut-down event; indeed, in the cases considered it appeared unlikely that scale failure would occur in any practically meaningful time due to strains accumulated during operation in a load-following mode in the absence of a shut down. The accuracy of the algorithms used for the kinetics of oxide growth appeared to be a very important consideration, especially for alloy TP347 for which large effects on oxide growth rate are known to occur with changes in alloy grain size and surface cold work.

36 MATERIALS SCIENCE↗

Report of the Topical Group on Particle Dark Matter for Snowmass 2021

This report summarizes the findings of the CF1 Topical Subgroup to Snowmass 2021, which was focused on particle dark matter. One of the most important scientific goals of the next decade is to reveal the nature of dark matter (DM). To accomplish this goal, we must delve deep, to cover high priority targets including weakly-interacting massive particles (WIMPs), and search wide, to explore as much motivated DM parameter space as possible. A diverse, continuous portfolio of experiments at large, medium, and small scales that includes both direct and indirect detection techniques maximizes the probability of discovering particle DM. Detailed calibrations and modeling of signal and background processes are required to make a convincing discovery. In the event that a candidate particle is found through different means, for example at a particle collider, the program described in this report is also essential to show that it is consistent with the actual cosmological DM. The US has a leading role in both direct and indirect detection dark matter experiments -- to maintain this leading role, it is imperative to continue funding major experiments and support a robust R&D program.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Scale sensitivity of ethanol production via consolidated bioprocessing with consideration of feedstock cost

We examine feedstock cost and minimum selling price for ethanol production from corn stover as a function of scale, stover yield, participation rate, and price incentives for two conversion technologies: a conventional base case featuring thermochemical pretreatment with added cellulase, and an advanced case featuring consolidated bioprocessing with cotreatment (C-CBP). Delivered feedstock cost ranged from $\$85$ Mg −1 at small (10 million gallons year −1 or ~38 million L year −1 ) scale with high yield and participation rates to $\$124$ Mg −1 at large scale (60 million gallons year −1 or 227 million L year −1 ) and low yield and participation rates. The minimum ethanol selling price (MESP) was approximately twofold lower for the advanced case compared with the base case. The payback period was several times lower for the advanced case compared with the base case, with increasing disparity at smaller scales, and was highly sensitive to ethanol price supports. For both C-CBP and the conventional processing paradigm, MESP decreased with increasing scale, indicating that the cost penalty due to higher feedstock transport distances was more than outweighed by lower capital costs. However, the cost penalty for operation at small scale, expressed in $ gallon −1 ethanol, is lower for C-CBP than for the conventional paradigm by roughly twofold. Particularly for initial applications of C-CBP, we speculate that this cost penalty will likely be modest compared with the anticipated benefits of small-scale operation such as increased opportunity to use existing infrastructure, easier plant siting and supply chain establishment, and lower total investment required.

biorefinery scale↗

ConnectX-2 InfiniBand Management Queues: First Investigation of the New Support for Network Offloaded Collective Operations

This paper introduces the newly developed Infini-Band (IB) Management Queue capability, used by the Host Channel Adapter (HCA) to manage network task data flow dependancies, and progress the communications associated with such flows. These tasks include sends, receives, and the newly supported wait task, and are scheduled by the HCA based on a data dependency description provided by the user. This functionality is supported by the ConnectX-2 HCA, and provides the means for delegating collective communication management and progress to the HCA, also known as collective communication offload. This provides a means for overlapping collective communications managed by the HCA and computation on the Central Processing Unit (CPU), thus making it possible to reduce the impact of system noise on parallel applications using collective operations. This paper further describes how this new capability can be used to implement scalable Message Passing Interface (MPI) collective operations, describing the high level details of how this new capability is used to implement the MPI Barrier collective operation, focusing on the latency sensitive performance aspects of this new capability. This paper concludes with small scale benchmark experiments comparing implementations of the barrier collective operation, using the new network offload capabilities, with established point-to-point based implementations of these same algorithms, which manage the data flow using the central processing unit. These early results demonstrate the promise this new capability provides to improve the scalability of high performance applications using collective communications. The latency of the HCA based implementation of the barrier is similar to that of the best performing point-to-point based implementation managed by the central processing unit, starting to outperform these as the number of processes involved in the collective operation increases.

Graham, Richard L↗

ConnectX2 In niBand Management Queues: New support for Network Of oaded

This paper introduces the newly developed InfiniBand (IB) Management Queue capability, used by the Host Channel Adapter (HCA) to manage network task data flow dependancies, and progress the communications associated with such flows. These tasks include sends, receives, and the newly supported wait task, and are scheduled by the HCA based on a data dependency description provided by the user. This functionality is supported by the ConnectX-2 HCA, and provides the means for delegating collective communication management and progress to the HCA, also known as collective communication offload. This provides a means for overlapping collective communications managed by the HCA and computation on the Central Processing Unit (CPU), thus making it possible to reduce the impact of system noise on parallel applications using collective operations. This paper further describes how this new capability can be used to implement scalable Message Passing Interface (MPI) collective operations, describing the high level details of how this new capability is used to implement the MPI Barrier collective operation, focusing on the latency sensitive performance aspects of this new capability. This paper concludes with small scale benchmark experiments comparing implementations of the barrier collective operation, using the new network offload capabilities, with established point-to-point based implementations of these same algorithms, which manage the data flow using the central processing unit. These early results demonstrate the promise this new capability provides to improve the scalability of high-performance applications using collective communications. The latency of the HCA based implementation of the barrier is similar to that of the best performing point-to-point based implementation managed by the central processing unit, starting to outperform these as the number of processes involved in the collective operation increases.

99 GENERAL AND MISCELLANEOUS↗

The Role of Small RNA in Biomass Deposition (Final Report for ER65039)

Our objective in this project was to discover the role of sRNA in regulating both biomass biosynthesis and perenniality in the Andropogoneae feedstock grasses. Our central hypothesis was that there is a time-and space specific sRNA network playing a crucial role in regulating processes associated with cell wall biosynthesis, flowering time control, overwintering/juvenility, and nutrient sequestration in the feedstock grasses. To address this, we performed a large scale biological project consisting of the growth of material, generation of Illumina libraries, sequencing and analysis for small RNA, mRNA and Degradome / cmRNA. Our subsidiary objectives included analysis of the biology of small RNAs and the cell wall composition of Miscanthus. These objectives have all been completed, one publication is in print, one is submitted and several more are in progress.

09 BIOMASS FUELS↗

Temporal Subsampling Diminishes Small Spatial Scales in Recurrent Neural Network Emulators of Geophysical Turbulence

The immense computational cost of traditional numerical weather and climate models has sparked the development of machine learning (ML) based emulators. Because ML methods benefit from long records of training data, it is common to use data sets that are temporally subsampled relative to the time steps required for the numerical integration of differential equations. Here, we investigate how this often overlooked processing step affects the quality of an emulator's predictions. We implement two ML architectures from a class of methods called reservoir computing: (a) a form of Nonlinear Vector Autoregression (NVAR), and (b) an Echo State Network (ESN). Despite their simplicity, it is well documented that these architectures excel at predicting low dimensional chaotic dynamics. We are therefore motivated to test these architectures in an idealized setting of predicting high dimensional geophysical turbulence as represented by Surface Quasi-Geostrophic dynamics. In all cases, subsampling the training data consistently leads to an increased bias at small spatial scales that resembles numerical diffusion. Interestingly, the NVAR architecture becomes unstable when the temporal resolution is increased, indicating that the polynomial based interactions are insufficient at capturing the detailed nonlinearities of the turbulent flow. The ESN architecture is found to be more robust, suggesting a benefit to the more expensive but more general structure. Spectral errors are reduced by including a penalty on the kinetic energy density spectrum during training, although the subsampling related errors persist. Future work is warranted to understand how the temporal resolution of training data affects other ML architectures.

58 GEOSCIENCES↗