Search NASA⌕ Search

SEARCH · Search NASA

Results for “Ensemble computing.”

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 415 records · Page 23

Correlating isothermal compressibility to nucleon fluctuations in the inner crust of neutron stars

The question of how and which physical observables or thermodynamic parameters can best predict the onset of a possible phase transition in the inner crust of neutron stars remains largely unresolved. Here, using semiclassical Monte Carlo simulations, we investigate the isothermal compressibility and density fluctuations in a region of relevance to the dynamics of the inner crust. We show that the isothermal compressibility serves as a robust observable to characterize the transition from the nonuniform crust to the uniform core for proton fractions over 0.2. Moreover, we show explicitly how the two-component isothermal compressibility, computed using the Kirkwood-Buff theory, is directly connected to the fluctuations in the number density, recorded in the grand canonical ensemble by monitoring the number of particles in a small volume located at the center of the simulation box. That is, we compute mean-square particle fluctuations and compare them against the isothermal compressibility for different proton fractions. Although our results show that the mean-square particle fluctuations are proportional to the isothermal compressibility, the lack of a perfect correlation is attributed to the relatively small number of particles included in the simulations. The nonunity slope observed in the dimensionless isothermal compressibility—total nucleon fluctuation variance relationship suggests that the inner crust of neutron stars is composed of anisotropic and inhomogeneous matter.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Fast Linked Analyses for Scenario-Based Hierarchies

Summary When using computer models to provide policy support it is normal to encounter ensembles that test only a handful of feasible or idealized decision scenarios. We present a new methodology for performing multilevel emulation of a complex model as a function of any decision within a predefined class that makes specific use of a scenario ensemble of opportunity on a fast or early version of a simulator and a small, well-chosen, design on our current simulator of interest. The method exploits a geometrical approach to Bayesian inference and is designed to be fast, to facilitate detailed diagnostic checking of our emulators by allowing us to carry out many analyses very quickly. Our motivating application involved constructing an emulator for the UK Met Office Hadley Centre coupled climate model HadCM3 as a function of carbon dioxide forcing, which was part of a ‘RAPID’ programme deliverable to the UK Met Office funded by the Natural Environment Research Council. Our application involved severe time pressure as well as limited access to runs of HadCM3 and a scenario ensemble of opportunity on a lower resolution version of the model.

Williamson, Daniel↗

Final Project Report

Most climate models are currently run with grid spacings of around 100km, which, with today’s computing power, allows for long (up to 1000 year) simulations, or ensembles of simulations to explore climate change and variability. However this grid spacing does not resolve important components of the weather/climate system such as atmospheric fronts and mesoscale systems, and ocean boundary currents and eddies. The overall aim of this project has been to look at the effect of these small-scale features on the weather/climate system using a suite of high and low resolution climate models, idealized models and observations. This project was only possible due to the highly scalable aspect of the CAM Spectral Element dynamical core, and the significant resources allocated at Yellowstone and NERSC for which we are grateful.

54 ENVIRONMENTAL SCIENCES↗

Collaborative project. Ocean-atmosphere interaction from meso- to planetary-scale. Mechanics, parameterization, and variability

Most climate models are currently run with grid spacings of around 100km, which, with today’s computing power, allows for long (up to 1000 year) simulations, or ensembles of simulations to explore climate change and variability. However this grid spacing does not resolve important components of the weather/climate system such as atmospheric fronts and mesoscale systems, and ocean boundary currents and eddies. The overall aim of this project has been to look at the effect of these small-scale features on the weather/climate system using a suite of high and low resolution climate models, idealized models and observations. High-resolution global coupled integrations using CAM/CESM were carried out at NCAR by the lead PI. At TAMU, we have complemented the work at NCAR by analyzing datasets from the high-resolution (28km) CESM integrations (Small et al., 2014) as well as very high resolution (9km, 3km) runs using a coupled regional climate (CRCM) carried out locally. The main tasks carried out were: 1. Analysis of surface wind in observations and high-resolution CAM/CCSM simulations 2. Development of a feature-tracking algorithm for studying midlatitude air-sea interaction by following oceanic mesoscale eddies and creating composites of the atmospheric response overlying the eddies. 3. Applying the Lagrangian analysis technique in the Gulf Stream region to compare data from observational reanalyses, global CESM coupled simulations, 9km regional coupled simulations and 3km convection-resolving regional coupled simulations. Our main findings are that oceanic mesoscale eddies influence not just the atmospheric boundary layer above them, but also the lower portions of the free troposphere above the boundary layer. Such a vertical response could have implications for a remote influence of Gulf Stream oceanic eddies on North Atlantic weather patterns through modulation of the storm track, similar to what has been noted in the North Pacific. The coarse resolution observational reanalyses perhaps underestimate the atmospheric response, but the 28km global model resolution appears to be adequate to capture some, but not all, aspects of the boundary response. The higher resolution regional models show a stronger response in certain fields such as the latent heat flux.

58 GEOSCIENCES↗

Terrain-Influenced Winds and Fire-Fire Interactions in Wildland Fire Simulations [Dissertation]

Ensemble-based approaches to prescribed fire planning cannot be supported by computational fluid dynamics based models like FIRETEC and the Wildland-Urban Interface Fire Dynamics Simulator (WFDS) because they are too computationally expensive and cannot leverage large eddy simulation approaches like CAWFE and WRF-SFIRE because they have too coarse of resolution. QUIC-Fire was developed to fill this gap but it cannot currently address complex terrain, that is typical for instance in the Western United States. This dissertation describes a variety of improvements made to QUIC-Fire and its various incorporated algorithms in an effort to make it a viable tool in simulating wildland and prescribed fires on terrain. The modifications made to QUIC-Fire are described in three chapters. The first chapter describes the extension of the diagnostic wind model QUIC-URB, the wind engine of QUIC-Fire, to a terrain-following coordinate system. The terraininfluenced winds it generates are analyzed and compared. In particular, this chapter presents the mathematical derivation of the wind solver leading to a linear system of equations that are solved through the successive over-relaxation method. The model is validated against a standard test used in previous works (the Askervein Hill) and against a new dataset from measurements in the Socorro Mountains, New Mexico. The terrain-following implementation captures the correct phenomenology for the isolated Askervein Hill, with a wind speed up at the top of the hill. The model agrees well with measurements on the upwind side of the peak, but overestimates speed-up on the downwind side of the hill. This is due to the inability of the model to generate flow separation and wake-eddy dynamics. On a common laptop, the divergence-free wind field is obtained in 6 s, making the solver appealing for coupled fire-atmosphere simulations. The Socorro Mountain is highly complex, with many cliff faces, peaks, and valleys. Although the model captures the magnitude and direction of inlet and outlet areas of the domain, it performs rather poorly in the valley region and in the regions near the steep cliffs. Hence, the model shows good agreement with data in areas of open sloped terrain but lacks in areas where flow separation and thermally driven effects may be present (neither effect is addressed in this work). In the second chapter the implementation of the terrain-following version of QUIC-URB into QUIC-Fire, and the necessary changes needed to include terrain are described. No changes to the underlying fire spread algorithm are made other than what is required to correctly account for the inclusion of terrain. Previously published FIRETEC results that use five different topographies that share the same centerline profile are compared to simulation results from the modified QUIC-Fire that use the same topographies and fuels. QUIC-Fire results show overall similar behaviors in terms of how the topographies affect fire shapes and trends in spread rates. Due to the terrain-following version of QUIC-URB being unable to generate flow separations at the crest of hills, fire spread rates in these regions across all topographies are over-predicted when compared to FIRETEC. Lateral fire growth shows similar trends with FIRETEC between topographies but does not capture the increase in spread due to a diagonal interface between grassland and forested fuel region of the domain. These results suggest that there are three algorithms within QUIC-Fire that could use improvement: how flame tilt angle is accounted for, the incorporation of non-local drag effects, and the inclusion of the wake-eddy parameterizations that are used in QUIC-URB. Lastly, the third chapter describes a modification to the initial guess used for the calculation of the QUIC-URB mass-conserved wind solution during fire simulations. The modification is aimed at improving fire-fire interactions in QUIC-Fire simulations. The modification consists of using the solution from the previous timestep as the starting point for the calculation of the solution for the next timestep. Fire-fire interactions is greatly improved by the change but a new source of error is introduced. Due to how plumes are modelled in QUIC-Fire the new solution contains errors where gaps in the plume structure are present. However, these errors are mostly limited to the upper atmosphere, where they do not affect fire behavior at the surface, and their magnitude isn’t significant enough to discount the amount of new fire phenomenology now captured in QUIC-Fire with the change.

58 GEOSCIENCES↗

The Decay Constants $f_{D_s}, f_{D+}, f_{B_s}$ and $f_B$ from Lattice QCD

In this report we summarize results for the leptonic decay constants $f_{D_s} , f_{D+}, f_{B_s}$ and $f_B$, obtained at the three lattice spacings a $\approx$ 0.09, 0.125 and 0.15 fm, from the now concluded first computational phase of our on-going project. The decay constants are computed on the MILC collaboration’s 2+1 flavor asqtad gauge ensembles. We use clover heavy quarks in the Fermilab interpretation and asqtad improved staggered light quarks. For each of the D and B meson systems, a simultaneous chiral and continuum extrapolation is performed with partially quenched lattice results. We show how improvements being implemented now, such as, higher statistics, additional finer lattice spacings and better determinations of inputs to the calculation, are expected to reduce future statistical and systematic uncertainties.

Simone, J.↗

The Ds and D+ Leptonic Decay Constants from Lattice QCD

We present the leptonic decay constants f{sub D{sub s}} and f{sub D{sup +}} computed on the MILC collaboration's 2+1 flavor asqtad gauge ensembles. We use clover heavy quarks with the Fermilab interpretation and improved staggered light quarks. The simultaneous chiral and continuum extrapolation, which determines both decay constants, includes partially-quenched lattice results at lattice spacings a {approx} 0.09, 0.12 and 0.15 fm. We have made several recent improvements in our analysis: (a) we include terms in the fit describing leading order heavy-quark discretization effects, (b) we have adopted a more precise input r{sub 1} value consistent with our other D and B meson studies, (c) we have retuned the input bare charm masses based upon the new r{sub 1}. Our preliminary results are f{sub D{sub s}} = 260 {+-} 10 MeV and f{sub D{sup +}} = 217 {+-} 10 MeV.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Scalable Comparative Visualization of Ensembles of Call Graphs

Optimizing the performance of large-scale parallel codes is critical for efficient utilization of computing resources. Code developers often explore various execution parameters, such as hardware configurations, system software choices, and application parameters, and are interested in detecting and understanding bottlenecks in different executions. They often collect hierarchical performance profiles represented as call graphs, which combine performance metrics with their execution contexts. The crucial task of exploring multiple call graphs together is tedious and challenging because of the many structural differences in the execution contexts and significant variability in the collected performance metrics (e.g., execution runtime). In this paper, we present Ensemble CallFlow to support the exploration of ensembles of call graphs using new types of visualizations, analysis, graph operations, and features. We introduce ensemble-Sankey , a new visual design that combines the strengths of resource-flow (Sankey) and box-plot visualization techniques. Whereas the resource-flow visualization can easily and intuitively describe the graphical nature of the call graph, the box plots overlaid on the nodes of Sankey convey the performance variability within the ensemble. Our interactive visual interface provides linked views to help explore ensembles of call graphs, e.g., by facilitating the analysis of structural differences, and identifying similar or distinct call graphs. Finally, we demonstrate the effectiveness and usefulness of our design through case studies on large-scale parallel codes.

97 MATHEMATICS AND COMPUTING↗

A localized ensemble of approximate Gaussian processes for fast sequential emulation

More attention has been given to the computational cost associated with the fitting of an emulator. Substantially less attention is given to the computational cost of using that emulator for prediction. This is primarily because the cost of fitting an emulator is usually far greater than that of obtaining a single prediction, and predictions can often be obtained in parallel. In many settings, especially those requiring Markov Chain Monte Carlo, predictions may arrive sequentially and parallelization is not possible. In this case, using an emulator procedure which can produce accurate predictions efficiently can lead to substantial time savings in practice. In this paper, we propose a global model approximate Gaussian process framework via extension of a popular local approximate Gaussian process (laGP) framework. Our proposed emulator can be viewed as a treed Gaussian process where the leaf nodes are laGP models, and the tree structure is learned greedily as a function of the prediction stream. The suggested method (called leapGP) has interpretable tuning parameters which control the time‐memory trade‐off. One reasonable choice of settings leads to an emulator with a training cost and makes predictions rapidly with an asymptotic amortized cost of .

97 MATHEMATICS AND COMPUTING↗

The Resolution Sensitivity of Northern Hemisphere Blocking in Four 25-km Atmospheric Global Circulation Models

The aim of this study is to investigate if the representation of Northern Hemisphere blocking is sensitive to resolution in current-generation atmospheric global circulation models (AGCMs). An evaluation is thus conducted of how well atmospheric blocking is represented in four AGCMs whose horizontal resolution is increased from a grid spacing of more than 100 km to about 25 km. It is shown that Euro-Atlantic blocking is simulated overall more credibly at higher resolution (i.e., in better agreement with a 50-yr reference blocking climatology created from the reanalyses ERA-40 and ERA-Interim). The improvement seen with resolution depends on the season and to some extent on the model considered. Euro-Atlantic blocking is simulated more realistically at higher resolution in winter, spring, and autumn, and robustly so across the model ensemble. The improvement in spring is larger than that in winter and autumn. Summer blocking is found to be better simulated at higher resolution by one model only, with little change seen in the other three models. The representation of Pacific blocking is not found to systematically depend on resolution. Despite the improvements seen with resolution, the 25-km models still exhibit large biases in Euro-Atlantic blocking. For example, three of the four 25-km models underestimate winter northern European blocking frequency by about one-third. The resolution sensitivity and biases in the simulated blocking are shown to be in part associated with the mean-state biases in the models' midlatitude circulation.

54 ENVIRONMENTAL SCIENCES↗

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↗

Coherence-Induced Deep Thermalization Transition in Random Permutation Quantum Dynamics

We report a phase transition in the projected ensemble—the collection of postmeasurement wave functions of a local subsystem obtained by measuring its complement. The transition emerges in systems undergoing random permutation dynamics, a type of quantum time evolution wherein computational basis states are shuffled without creating superpositions. It separates a phase exhibiting deep thermalization, where the projected ensemble is distributed over Hilbert space in a maximally entropic fashion (Haar random), from a phase where it is minimally entropic (“classical bit-string ensemble”). Crucially, this deep thermalization transition is invisible to the subsystem’s density matrix, which always exhibits thermalization to infinite temperature across the phase diagram. Through a combination of analytical arguments and numerical simulations, we show that the transition is tuned by the total amount of injected by the input state and the measurement basis, and is exhibited robustly across different microscopic models. Our findings represent a novel form of ergodicity-breaking universality in quantum many-body dynamics, characterized not by a failure of regular thermalization, but rather by a failure of deep thermalization.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Arrested-motility states in populations of shape-anisotropic active Janus particles

The emergence of large-scale collective phenomena from simple interactions between individual units is a hallmark of active matter systems. Active colloids with alignment-dominated interparticle interactions tend to develop orientational order and form motile coherent states, such as flocks and swarms. Alternatively, a combination of self-propulsion and excluded-volume interactions results in self-trapping and active phase separation into dense clusters. Here, we reveal unconventional arrested-motility states in ensembles of active discoidal particles powered by induced-charge electrophoresis. Combining experiments and computational modeling, we demonstrate that the shape asymmetry of the particles promotes the hydrodynamically assisted formation of active particles’ bound states in a certain range of excitation parameters, ultimately leading to a spontaneous collective state with arrested motility. Unlike the jammed clusters obtained through self-trapping, the arrested-motility phase remains sparse, dynamic, and reconfigurable. The demonstrated mechanism of phase separation seeded by bound state formation in ensembles of oblate active particles is generic and should be applicable to other active colloidal systems.

42 ENGINEERING↗

Ensemble-based enzyme design can recapitulate the effects of laboratory directed evolution in silico

The creation of artificial enzymes is a key objective of computational protein design. Although de novo enzymes have been successfully designed, these exhibit low catalytic efficiencies, requiring directed evolution to improve activity. Here, we use room-temperature X-ray crystallography to study changes in the conformational ensemble during evolution of the designed Kemp eliminase HG3 (k cat /K M 146 M -1 s -1 ). We observe that catalytic residues are increasingly rigidified, the active site becomes better pre-organized, and its entrance is widened. Based on these observations, we engineer HG4, an efficient biocatalyst (k cat /K M 103,000 M -1 s -1 ) containing key first and second-shell mutations found during evolution. HG4 structures reveal that its active site is pre-organized and rigidified for efficient catalysis. Our results show how directed evolution circumvents challenges inherent to enzyme design by shifting conformational ensembles to favor catalytically-productive sub-states, and suggest improvements to the design methodology that incorporate ensemble modeling of crystallographic data.

59 BASIC BIOLOGICAL SCIENCES↗

Application of High Performance Computing for Studying Cyclic Variability in Dilute Internal Combustion Engines

Combustion instabilities in dilute internal combustion engines are manifest in cyclic variability (CV) in engine performance measures such as integrated heat release or shaft work. Understanding the factors leading to CV is important in model-based control, especially with high dilution where experimental studies have demonstrated that deterministic effects can become more prominent. Observation of enough consecutive engine cycles for significant statistical analysis is standard in experimental studies but is largely wanting in numerical simulations because of the computational time required to compute hundreds or thousands of consecutive cycles. We have proposed and begun implementation of an alternative approach to allow rapid simulation of long series of engine dynamics based on a low-dimensional mapping of ensembles of single-cycle simulations which map input parameters to output engine performance. This paper details the use Titan at the Oak Ridge Leadership Computing Facility to investigate CV in a gasoline direct-injected spark-ignited engine with a moderately high rate of dilution achieved through external exhaust gas recirculation. The CONVERGE™ CFD software was used to perform single-cycle simulations with imposed variations of operating parameters and boundary conditions selected according to a sparse grid sampling of the parameter space. Using an uncertainty quantification technique, the sampling scheme is chosen similar to a design of experiments grid but uses algorithms designed to minimize the number of samples required to achieve a desired degree of accuracy. The simulations map input parameters to output metrics of engine performance for a single cycle, and by mapping over a large parameter space, results can be interpolated from within that space. This interpolation scheme forms the basis for a low-dimensional ‘metamodel’ (or model of a model) which can be used to mimic the dynamical behavior of corresponding high-dimensional simulations. Simulations of high-EGR spark-ignition combustion cycles within a parametric sampling grid were performed and analyzed statistically, and sensitivities of the physical factors leading to high CV are presented. With these results, the prospect of producing low-dimensional metamodels to describe engine dynamics at any point in the parameter space will be discussed. Additionally, modifications to the methodology to account for nondeterministic effects in the numerical solution environment are proposed.

Finney, Charles E. A.↗

DART-PFLOTRAN: An ensemble-based data assimilation system for estimating subsurface flow and transport model parameters

Ensemble-based Data Assimilation (EDA), based on the Monte Carlo approach, has been effectively applied to estimate model parameters through inverse modeling in subsurface flow and transport problems. However, implementation of EDA approach involves a complicated workflow that include setting up and executing ensemble forward model simulations, processing observations and model simulation results for parameter updates, and repeat for sequential or iterative EDA. To facilitate the management of such workflow and lower the barriers for adopting EDA-based parameter estimation in subsurface science, we develop a generic software frame-work linking the Data Assimilation Research Testbed (DART) with a massively parallel subsurface FLOw and TRANsport code PFLOTRAN. The new DART-PFLOTRAN leverages both the core data assimilation engines in DART and the computational power afforded by PFLOTRAN. In addition to the standard smoother and filtering options, DART-PFLOTRAN enables an iterative EDA workflow based on the Ensemble Smoother for Multiple Data Assimilation method (ES-MDA) to improve estimation accuracy for nonlinear forward problems. Here, we verify the implementation of ES-MDA in DART-PFLOTRAN using two synthetic cases designed to estimate static permeability and dynamic exchange fluxes across the riverbed, respectively, from continuous temperature measurements made across a depth profile. One-dimensional hydro-thermal simulations are performed in both cases to relate temperature responses with the parameters of interest. In the case of estimating dynamic parameters, we demonstrate the flexibility of DART-PFLOTRAN in automating sequential ES-MDA workflow, which will significantly reduce the time researchers spend on managing complex workflows in similar applications. Both studies yield accurate estimations of the parameters compared to their synthetic truth, while ES-MDA leads to more accurate estimation when a high level of nonlinearity exist between observed responses and unknown parameters. With a code base in Python and Fortran, DART-PFLOTRAN paves the way for applications in large-scale subsurface inverse modeling by automating the complex workflow of sequential ES-MDA that can be executed on various computing platforms.

97 MATHEMATICS AND COMPUTING↗

Forces and stress in second order Møller-Plesset perturbation theory for condensed phase systems within the resolution-of-identity Gaussian and plane waves approach

The forces acting on the atoms as well as the stress tensor are crucial ingredients for calculating the structural and dynamical properties of systems in the condensed phase. Here, these derivatives of the total energy are evaluated for the second-order Møller-Plesset perturbation energy (MP2) in the framework of the resolution of identity Gaussian and plane waves method, in a way that is fully consistent with how the total energy is computed. This consistency is non-trivial, given the different ways employed to compute Coulomb, exchange, and canonical four center integrals, and allows, for example, for energy conserving dynamics in various ensembles. Based on this formalism, a massively parallel algorithm has been developed for finite and extended system. The designed parallel algorithm displays, with respect to the system size, cubic, quartic, and quintic requirements, respectively, for the memory, communication, and computation. All these requirements are reduced with an increasing number of processes, and the measured performance shows excellent parallel scalability and efficiency up to thousands of nodes. Additionally, the computationally more demanding quintic scaling steps can be accelerated by employing graphics processing units (GPU’s) showing, for large systems, a gain of almost a factor two compared to the standard central processing unit-only case. In this way, the evaluation of the derivatives of the RI-MP2 energy can be performed within a few minutes for systems containing hundreds of atoms and thousands of basis functions. With good time to solution, the implementation thus opens the possibility to perform molecular dynamics (MD) simulations in various ensembles (microcanonical ensemble and isobaric-isothermal ensemble) at the MP2 level of theory. Geometry optimization, full cell relaxation, and energy conserving MD simulations have been performed for a variety of molecular crystals including NH{sub 3}, CO{sub 2}, formic acid, and benzene.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Scalable deep learning for watershed model calibration

Watershed models such as the Soil and Water Assessment Tool (SWAT) consist of high-dimensional physical and empirical parameters. These parameters often need to be estimated/calibrated through inverse modeling to produce reliable predictions on hydrological fluxes and states. Existing parameter estimation methods can be time consuming, inefficient, and computationally expensive for high-dimensional problems. In this paper, we present an accurate and robust method to calibrate the SWAT model (i.e., 20 parameters) using scalable deep learning (DL). We developed inverse models based on convolutional neural networks (CNN) to assimilate observed streamflow data and estimate the SWAT model parameters. Scalable hyperparameter tuning is performed using high-performance computing resources to identify the top 50 optimal neural network architectures. We used ensemble SWAT simulations to train, validate, and test the CNN models. We estimated the parameters of the SWAT model using observed streamflow data and assessed the impact of measurement errors on SWAT model calibration. We tested and validated the proposed scalable DL methodology on the American River Watershed, located in the Pacific Northwest-based Yakima River basin. Our results show that the CNN-based calibration is better than two popular parameter estimation methods (i.e., the generalized likelihood uncertainty estimation [GLUE] and the dynamically dimensioned search [DDS], which is a global optimization algorithm). For the set of parameters that are sensitive to the observations, our proposed method yields narrower ranges than the GLUE method but broader ranges than values produced using the DDS method within the sampling range even under high relative observational errors. The SWAT model calibration performance using the CNNs, GLUE, and DDS methods are compared using R 2 and a set of efficiency metrics, including Nash-Sutcliffe, logarithmic Nash-Sutcliffe, Kling-Gupta, modified Kling-Gupta, and non-parametric Kling-Gupta scores, computed on the observed and simulated watershed responses. The best CNN-based calibrated set has scores of 0.71, 0.75, 0.85, 0.85, 0.86, and 0.91. The best DDS-based calibrated set has scores of 0.62, 0.69, 0.8, 0.77, 0.79, and 0.82. The best GLUE-based calibrated set has scores of 0.56, 0.58, 0.71, 0.7, 0.71, and 0.8. The scores above show that the CNN-based calibration leads to more accurate low and high streamflow predictions than the GLUE and DDS sets. Our research demonstrates that the proposed method has high potential to improve our current practice in calibrating large-scale integrated hydrologic models.

54 ENVIRONMENTAL SCIENCES↗