SEARCH · Engineering Papers

Results for “model parameterization”

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.

SourceResearch connectionFrom yearThrough year

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

At least 1,225 records · Page 68

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Self-similar Reynolds-averaged mechanical–scalar turbulence models for Rayleigh–Taylor mixing induced by power-law accelerations in the small Atwood number limit

Analytical self-similar solutions to two-, three-, and four-equation Reynolds-averaged mechanical–scalar turbulence models describing turbulent Rayleigh–Taylor mixing driven by a temporal power-law acceleration are derived in the small Atwood number (Boussinesq) limit. The solutions generalize those previously derived for constant acceleration Rayleigh–Taylor mixing for models based on the turbulent kinetic energy K and its dissipation rate ε, together with the scalar variance S and its dissipation rate χ [O. Schilling, “Self-similar Reynolds-averaged mechanical–scalar turbulence models for Rayleigh–Taylor, Richtmyer–Meshkov, and Kelvin–Helmholtz instability-induced mixing in the small Atwood number limit,” Phys. Fluids 33, 085129 (2021)]. The turbulent fields are expressed in terms of the model coefficients and power-law exponent, with their temporal power-law scalings obtained by requiring that the self-similar equations are explicitly time-independent. Mixing layer growth parameters and other physical observables are obtained explicitly as functions of the model coefficients and parameterized by the exponent of the power-law acceleration. Values for physical observables in the constant acceleration case are used to calibrate the two-, three-, and four-equation models, such that the self-similar solutions are consistent with experimental and numerical simulation data corresponding to a canonical (i.e., constant acceleration) Rayleigh–Taylor turbulent flow. The calibrated four-equation model is then used to numerically reconstruct the mean and turbulent fields, and turbulent equation budgets across the mixing layer for several values of the power-law exponent. Finally, the reference solutions derived here can be used to understand the model predictions for strongly accelerated or decelerated Rayleigh–Taylor mixing in the large Reynolds number limit.

42 ENGINEERING↗

DOE OSTI2019

Evaluating Climate Model Simulations of the Radiative Forcing and Radiative Response at Earth’s Surface

In this work, we analyze the radiative forcing and radiative response at Earth’s surface, where perturbations in the radiation budget regulate the atmospheric hydrological cycle. By applying a radiative kernel-regression technique to CMIP5 climate model simulations where CO 2 is instantaneously quadrupled, we evaluate the intermodel spread in surface instantaneous radiative forcing, radiative adjustments to this forcing, and radiative responses to surface warming. The cloud radiative adjustment to CO 2 forcing and the temperature-mediated cloud radiative response exhibit significant intermodel spread. In contrast to its counterpart at the top of the atmosphere, the temperature-mediated cloud radiative response at the surface is found to be positive in some models and negative in others. Also, the compensation between the temperature-mediated lapse rate and water vapor radiative responses found in top-of-atmosphere calculations is not present for surface radiative flux changes. Instantaneous radiative forcing at the surface is rarely reported for model simulations; as a result, intermodel differences have not previously been evaluated in global climate models. We demonstrate that the instantaneous radiative forcing is the largest contributor to intermodel spread in effective radiative forcing at the surface. We also find evidence of differences in radiative parameterizations in current models and argue that this is a significant, but largely overlooked, source of bias in climate change simulations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2015

Multiscale Simulation of Moist Global Atmospheric Flows

The overarching goal of this award was to include phase changes of the water substance and accompanying latent heating and precipitation processes into the all-scale nonhydrostatic atmospheric dynamics EUlerian/LAGrangian (EULAG) model. The model includes fluid flow solver that is based on either an unabbreviated set of the governing equations (i.e., compressible dynamics) or a simplified set of equations without sound waves (i.e., sound-proof, either anelastic or pseudo-incompressible). The latter set has been used in small-scale dynamics for decades, but its application to the all-scale dynamics (from small-scale to planetary) has never been studied in practical implementations. The highlight of the project is the development of the moist implicit compressible model that can be run by applying time steps, as long as the anelastic model is limited only by the computational stability of the fluid flow and not by the speed of sound waves that limit the stability of explicit compressible models. Applying various versions of the EULAG model within the same numerical framework allows for an unprecedented comparison of solutions obtained with various sets of the governing equations and straightforward evaluation of the impact of various physical parameterizations on the model solutions. The main outcomes of this study are reported in three papers, two published and one currently under review. These papers include comparisons between model solutions for idealized moist problems across the range of scales from small to planetary. These tests include: moist thermals rising in the stable-stratified environment (following Grabowski and Clark, J. Atmos. Sci. 1991) and in the moist-neutral environment (after Bryan and Fritsch, Mon. Wea. Rev. 2002), moist flows over a mesoscale topography (as in Grabowski and Smolarkiewicz, Mon. Wea. Rev. 2002), deep convection in a sheared environment (following Weisman and Klemp, Mon. Wea. Rev. 1982), moist extension of the baroclinic wave on the sphere of Jablonowski and Williamson (Q. J. R. Met. Soc. 2006), and moist extension of the Held-Suarez idealized climate benchmark (Held and Suarez, Bull. Amer. Met. Soc., 1994).

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1998

Geometry Modeling and Grid Generation for Design and Optimization

Geometry modeling and grid generation (GMGG) have played and will continue to play an important role in computational aerosciences. During the past two decades, tremendous progress has occurred in GMGG; however, GMGG is still the biggest bottleneck to routine applications for complicated Computational Fluid Dynamics (CFD) and Computational Structures Mechanics (CSM) models for analysis, design, and optimization. We are still far from incorporating GMGG tools in a design and optimization environment for complicated configurations. It is still a challenging task to parameterize an existing model in today's Computer-Aided Design (CAD) systems, and the models created are not always good enough for automatic grid generation tools. Designers may believe their models are complete and accurate, but unseen imperfections (e.g., gaps, unwanted wiggles, free edges, slivers, and transition cracks) often cause problems in gridding for CSM and CFD. Despite many advances in grid generation, the process is still the most labor-intensive and time-consuming part of the computational aerosciences for analysis, design, and optimization. In an ideal design environment, a design engineer would use a parametric model to evaluate alternative designs effortlessly and optimize an existing design for a new set of design objectives and constraints. For this ideal environment to be realized, the GMGG tools must have the following characteristics: (1) be automated, (2) provide consistent geometry across all disciplines, (3) be parametric, and (4) provide sensitivity derivatives. This paper will review the status of GMGG for analysis, design, and optimization processes, and it will focus on some emerging ideas that will advance the GMGG toward the ideal design environment.

Samareh, Jamshid A.↗

DOE OSTI2023

On the influence of over-parameterization in manifold based surrogates and deep neural operators

Constructing accurate and generalizable approximators (surrogate models) for complex physico-chemical processes exhibiting highly non-smooth dynamics is challenging. The main question is what type of surrogate models we should construct and should these models be under-parameterized or over-parameterized. In this work, we propose new developments and perform comparisons for two promising approaches: manifold-based polynomial chaos expansion (m-PCE) and the deep neural operator (DeepONet), and we examine the effect of over-parameterization on generalization. While m-PCE enables the construction of a mapping by first identifying low-dimensional embeddings of the input functions, parameters, and quantities of interest (QoIs), a neural operator learns the nonlinear mapping via the use of deep neural networks. Here, we demonstrate the performance of these methods in terms of generalization accuracy by solving the 2D time-dependent Brusselator reaction-diffusion system with uncertainty sources, modeling an autocatalytic chemical reaction between two species. We first propose an extension of the m-PCE by constructing a mapping between latent spaces formed by two separate embeddings of the input functions and the output QoIs. To further enhance the accuracy of the DeepONet, we introduce weight self-adaptivity in the loss function. We demonstrate that the performance of m-PCE and DeepONet is comparable for cases of relatively smooth input-output mappings. However, when highly non-smooth dynamics is considered, DeepONet shows higher approximation accuracy. We also find that for m-PCE, modest over-parameterization leads to better generalization, both within and outside of distribution, whereas aggressive over-parameterization leads to over-fitting. In contrast, an even highly over-parameterized DeepONet leads to better generalization for both smooth and non-smooth dynamics. Furthermore, we compare the performance of the above models with another recently proposed operator learning model, the Fourier Neural Operator, and show that its over-parameterization also leads to better generalization. Taken together, our studies show that m-PCE can provide very good accuracy at very low training cost, whereas a highly over-parameterized DeepONet can provide better accuracy and robustness to noise but at higher training cost. In both methods, the inference cost is negligible.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2001

A Study of the Response of Deep Tropical Clouds to Mesoscale Processes: Modeling Strategies and Simulations of TOGA-COARE Convective Systems - Part 1

Interactions between deep tropical clouds over the western Pacific warm pool and the larger-scale environment are key to understanding climate change. Cloud models are an extremely useful tool in simulating and providing statistical information on heat and moisture transfer processes between cloud systems and the environment, and can therefore be utilized to substantially improve cloud parameterizations in climate models. In this paper, the Goddard Cumulus Ensemble (GCE) cloud-resolving model is used in multi-day simulations of deep tropical convective activity over the Tropical Ocean-Global Atmosphere Coupled Ocean-Atmosphere Response Experiment (TOGA COARE). Large-scale temperature and moisture advective tendencies, and horizontal momentum from the TOGA-COARE Intensive Flux Array (IFA) region, are applied to the GCE version which incorporates cyclical boundary conditions. Sensitivity experiments show that grid domain size produces the largest response to domain-mean temperature and moisture deviations, as well as cloudiness, when compared to grid horizontal or vertical resolution, and advection scheme. It is found that a minimum grid-domain size of 500 km is needed to adequately resolve the convective cloud features. The control experiment shows that the atmospheric heating and moistening is primarily a response to cloud latent processes of condensation/evaporation, and deposition/sublimation, and to a lesser extent, melting of ice particles. Air-sea exchange of heat and moisture is found to be significant, but of secondary importance, while the radiational response is small. The simulated rainfall and atmospheric heating and moistening, agrees well with observations, and performs favorably to other models simulating this case.

Johnson, Daniel E.↗

DOE OSTI2024

Bimodality in simulated precipitation frequency distributions and its relationship with convective parameterizations

Abstract Bimodality in precipitation frequency distributions is often evident in atmospheric models, but rarely in observations. This study i) proposes a metric to objectively quantify the bimodality in precipitation distributions, ii) evaluates model simulations contributed to the Coupled Model Intercomparison Project (CMIP) phase 5 (CMIP5), phase 6 (CMIP6), and the DYnamics of the Atmospheric general circulation Modeled On Non-hydrostatic Domains (DYAMOND) project by comparing them to satellite-based and reanalysis precipitation products, and iii) investigates possible origins of bimodal precipitation distributions. Our results reveal that about 83% (20 out of 24) of CMIP5 and 70% (21 out of 30) of CMIP6 models used in this study exhibit bimodal distributions. The few DYAMOND models that use a deep convective parameterization also show bimodal distributions, while most DYAMOND models do not. Predictably, the bimodality originates from the separation of precipitation process between resolved grid-scale and parameterized subgrid-scale. However, in a larger number of models bimodality arises from the parameterized subgrid-scale convective precipitation alone.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1991

Nonlinear influence of mesoscale land use on weather and climate

It is shown that the influence of mesoscale landscape spatial variability on the atmosphere must be parameterized or explicitly modeled in larger-scale atmospheric model simulations including general circulation models. The mesoscale fluxes of heat that result from this variability are shown to be of the same order of magnitude but with a different vertical structure than found for the turbulent fluxes. These conclusions are based on experiments in which no phase changes of water were permitted. To parameterize surface thermal inhomogeneities, the influence of landscape must be evaluated using spectral analysis or an equivalent procedure. To include the nonlinear contribution of each scale, numerical model simulations for the range of observed surface and overlying atmospheric conditions must be performed.

Pielke, R. A.↗

NASA NTRS1986

The effect of breaking gravity waves on the dynamics and chemistry of the mesosphere and lower thermosphere (invited review)

The influence of breaking gravity waves on the dynamics and chemical composition of the 60 to 110 km region is investigated with a two dimensional model that includes a parameterization of gravity wave momentum deposition and diffusion. The dynamical model is described by Garcia and Solomon (1983) and Solomon and Garcia (1983) and includes a complete chemical scheme for the mesosphere and lower thermosphere. The parameterization of Lindzen (1981) is used to calculate the momentum deposited and the turbulent diffusion produced by the gravity waves. It is found that wave momentum deposition drives a very vigorous mean meridional circulation, produces a very cold summer mesopause and reverse the zonal wind jets above about 85 km. The seasonal variation of the turbulent diffusion coefficient is consistent with the behavior of mesospheric turbulences inferred from MST radar echoes. The large degree of consistency between model results and various types of dynamical and chemical data supports very strongly the hypothesis that breaking gravity waves play a major role in determining the zonally-averaged dynamical and chemical structure of the 60 to 110 km region of the atmosphere.

Garcia, R. R.↗

DOE OSTI2014

Increasing water use efficiency along the C3 to C4 evolutionary pathway: a stomatal optimization perspective

C4 photosynthesis evolved independently numerous times, probably in response to declining atmospheric CO2 concentrations, but also to high temperatures and aridity, which enhance water losses through transpiration. Here, the environmental factors controlling stomatal behaviour of leaf-level carbon and water exchange were examined across the evolutionary continuum from C3 to C4 photosynthesis at current (400 μmol mol–1) and low (280 μmol mol–1) atmospheric CO2 conditions. To this aim, a stomatal optimization model was further developed to describe the evolutionary continuum from C3 to C4 species within a unified framework. Data on C3, three categories of C3–C4 intermediates, and C4 Flaveria species were used to parameterize the stomatal model, including parameters for the marginal water use efficiency and the efficiency of the CO2-concentrating mechanism (or C4 pump); these two parameters are interpreted as traits reflecting the stomatal and photosynthetic adjustments during the C3 to C4 transformation. Neither the marginal water use efficiency nor the C4 pump strength changed significantly from C3 to early C3–C4 intermediate stages, but both traits significantly increased between early C3–C4 intermediates and the C4-like intermediates with an operational C4 cycle. At low CO2, net photosynthetic rates showed continuous increases from a C3 state, across the intermediates and towards C4 photosynthesis, but only C4-like intermediates and C4 species (with an operational C4 cycle) had higher water use efficiencies than C3 Flaveria. The results demonstrate that both the marginal water use efficiency and the C4 pump strength increase in C4 Flaveria to improve their photosynthesis and water use efficiency compared with C3 species. These findings emphasize that the advantage of the early intermediate stages is predominantly carbon based, not water related.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1985

The Characterization of Surface Variegation Effects on Remote Sensing

Improvements in remote sensing capabilities hinge very directly upon attaining an understanding of the physical processes contributing to the measurements. In order to devise new measurement strategies and to learn better techniques for processing remotely gathered data, it is necessary to understand and to characterize the complex radiative interactions of the atmosphere-surface system. In particular, it is important to understand the role of atmospheric structure, ground reflectance inhomogeneity and ground bidirectional reflectance type. The goals, then, are to model, analyze, and parameterize the observable effects of three dimensional atmospheric structure and composition and two dimensional variations in ground albedo and bidirectional reflectance. To achieve these goals, a Monte Carlo radiative transfer code is employed to model and analyze the effects of many of the complications which are present in nature.

Pearce, W. A.↗

NASA NTRS2002

The Influence of Tropospheric Processes in Modeling the Middle Atmosphere with Gravity Waves

Our Numerical Spectral Model (NSM) extends from the ground up into the thermosphere and has a vertical grid point resolution of about 0.5 km to resolve the interactions of gravity waves (GWs) described with Hines' Doppler Spread Parameterization (DSP). This model produces in the stratosphere and mesosphere the major features of QBO, SAO, tides, and planetary waves. The purpose of this paper is to discuss results from an initial study with our 3D model that shows how certain tropospheric processes can affect the dynamics of the middle atmosphere. Under the influence of tropospheric heating, and augmented by GW interactions, two distinct but related processes can be identified. (1) A meridional circulation develops in the stratosphere, with rising motions at low latitudes that are in magnitude comparable to the downward propagation of the QBO. As Dunkerton pointed out, a larger GW source is then required to reproduce the observed QBO, which tends to move us closer to the values recommended for the DSP. This has significant consequences for our model results that describe the upper mesosphere, considering the general importance of GWs for this region and in influencing planetary waves (e.g., 2-day wave) and tides in particular. (2) Tropospheric heating produces zonal jets near the tropopause that are related to latitudinal variations in pressure and reversing temperature variations (resembling the dynamical conditions near the mesopause), which in turn is conducive to generate baroclinic instability. Modeling results show that our ability to generate the QBO critically depends on the magnitude of the temperature reversal that is a measure of this instability. Planetary waves are generated in this process, which can apparently interfere with or augment the GW interactions. As originally demonstrated by Lindzen and Holton, the eastward propagating Kelvin waves and westward propagating Rossby gravity waves (generated by tropospheric convection) can in principle provide the acceleration to influence the QBO, and we were able to confirm this with our 3D model.

Mayr, H. G.↗

DOE OSTI2015

Delineating Hydrofacies Spatial Distribution by Integrating Ensemble Data Assimilation and Indicator Geostatistics

This study develops a new framework of facies-based data assimilation for characterizing spatial distribution of hydrofacies and estimating their associated hydraulic properties. This framework couples ensemble data assimilation with transition probability-based geostatistical model via a parameterization based on a level set function. The nature of ensemble data assimilation makes the framework efficient and flexible to be integrated with various types of observation data. The transition probability-based geostatistical model keeps the updated hydrofacies distributions under geological constrains. The framework is illustrated by using a two-dimensional synthetic study that estimates hydrofacies spatial distribution and permeability in each hydrofacies from transient head data. Our results show that the proposed framework can characterize hydrofacies distribution and associated permeability with adequate accuracy even with limited direct measurements of hydrofacies. Our study provides a promising starting point for hydrofacies delineation in complex real problems.

Song, Xuehang↗

DOE OSTI2023

Multi-reward reinforcement learning based development of inter-atomic potential models for silica

Abstract Silica is an abundant and technologically attractive material. Due to the structural complexities of silica polymorphs coupled with subtle differences in Si–O bonding characteristics, the development of accurate models to predict the structure, energetics and properties of silica polymorphs remain challenging. Current models for silica range from computationally efficient Buckingham formalisms (BKS, CHIK, Soules) to reactive (ReaxFF) and more recent machine-learned potentials that are flexible but computationally costly. Here, we introduce an improved formalism and parameterization of BKS model via a multireward reinforcement learning (RL) using an experimental training dataset. Our model concurrently captures the structure, energetics, density, equation of state, and elastic constants of quartz (equilibrium) as well as 20 other metastable silica polymorphs. We also assess its ability in capturing amorphous properties and highlight the limitations of the BKS-type functional forms in simultaneously capturing crystal and amorphous properties. We demonstrate ways to improve model flexibility and introduce a flexible formalism, machine-learned ML-BKS, that outperforms existing empirical models and is on-par with the recently developed 50 to 100 times more expensive Gaussian approximation potential (GAP) in capturing the experimental structure and properties of silica polymorphs and amorphous silica.

36 MATERIALS SCIENCE↗

NASA NTRS1996

Two-stage model of radon-induced malignant lung tumors in rats: effects of cell killing

A two-stage stochastic model of carcinogenesis is used to analyze lung tumor incidence in 3750 rats exposed to varying regimens of radon carried on a constant-concentration uranium ore dust aerosol. New to this analysis is the parameterization of the model such that cell killing by the alpha particles could be included. The model contains parameters characterizing the rate of the first mutation, the net proliferation rate of initiated cells, the ratio of the rates of cell loss (cell killing plus differentiation) and cell division, and the lag time between the appearance of the first malignant cell and the tumor. Data analysis was by standard maximum likelihood estimation techniques. Results indicate that the rate of the first mutation is dependent on radon and consistent with in vitro rates measured experimentally, and that the rate of the second mutation is not dependent on radon. An initial sharp rise in the net proliferation rate of initiated cell was found with increasing exposure rate (denoted model I), which leads to an unrealistically high cell-killing coefficient. A second model (model II) was studied, in which the initial rise was attributed to promotion via a step function, implying that it is due not to radon but to the uranium ore dust. This model resulted in values for the cell-killing coefficient consistent with those found for in vitro cells. An "inverse dose-rate" effect is seen, i.e. an increase in the lifetime probability of tumor with a decrease in exposure rate. This is attributed in large part to promotion of intermediate lesions. Since model II is preferable on biological grounds (it yields a plausible cell-killing coefficient), such as uranium ore dust. This analysis presents evidence that a two-stage model describes the data adequately and generates hypotheses regarding the mechanism of radon-induced carcinogenesis.

NASA Program Radiation Health↗

NASA NTRS2011

Understanding Effective Diameter and Its Application to Terrestrial Radiation in Ice Clouds

The cloud property known as "effective diameter" or "effective radius", which in essence is the cloud particle size distribution (PSD) volume at bulk density divided by its projected area, is used extensively in atmospheric radiation transfer, climate modeling and remote sensing. This derives from the assumption that PSD optical properties can be uniquely described in terms of their effective diameter, D(sub e), and their cloud water content (CWC), henceforth referred to as the D(sub e)-CWC assumption. This study challenges this assumption, showing that while the D(sub e)-CWC assumption appears generally valid for liquid water clouds, it appears less valid for ice clouds in regions where (1) absorption is not primarily a function of either the PSD ice water content (IWC) or the PSD projected area, and (2) where wave resonance (i.e. photon tunneling) contributes significantly to absorption. These two regions often strongly coincide at terrestrial wavelengths when De less than ~60 m, which is where this D(sub e)-CWC assumption appears poorest. Treating optical properties solely in terms of D(sub e) and IWC may lead to errors up to 24%, 26% and 20% for terrestrial radiation in the window region regarding the absorption and extinction coefficients and the single scattering albedo, respectively. Outside the window region, errors may reach 33% and 42% regarding absorption and extinction. The magnitude and sign of these errors can change rapidly with wavelength, which may produce significant errors in climate modeling, remote sensing and other applications concerned with the wavelength dependence of radiation. Where the D(sub e)-CWC assumption breaks down, ice cloud optical properties appear to depend on D(sub e), IWC and the PSD shape. Optical property parameterizations in climate models and remote sensing algorithms based on historical PSD measurements may exhibit errors due to previously unknown PSD errors (i.e. the presence of ice artifacts due to the shattering of larger ice particles on the probe inlet tube during sampling). More recently developed cloud probes are designed to mitigate this shattering problem. Using realistic PSD shapes for a given temperature (and/or IWC) and cloud type may minimize errors associated with PSD shape in ice optics parameterizations and remote sensing algorithms. While this topic was investigated using two ice optics schemes (the Yang et al., 2005 database and the modified anomalous diffraction approximation, or MADA), a physical understanding of the limitations of the D(sub e)-IWC assumption was made possible by using MADA. MADA allows one to approximate the contribution of photon tunneling to absorption relative to other optical processes, which reveals that part of the error regarding the D(sub e)-IWC assumption can be associated with tunneling. By relating the remaining error to the radiation penetration depth in bulk ice (DELTA L) due to absorption, the domain where the D(sub e)-IWC assumption is weakest was described in terms of D(sub e) and DELTA L.

Mitchell, D. L.↗

DOE OSTI2020

Nonlinear sparse Bayesian learning for physics-based models

This paper addresses the issue of overfitting while calibrating unknown parameters of over-parameterized physics-based models with noisy and incomplete observations. Here, a semi-analytical Bayesian framework of nonlinear sparse Bayesian learning (NSBL) is proposed to identify sparsity among model parameters during Bayesian inversion. NSBL offers significant advantages over machine learning algorithm of sparse Bayesian learning (SBL) for physics-based models, such as 1) the likelihood function or the posterior parameter distribution is not required to be Gaussian, and 2) prior parameter knowledge is incorporated into sparse learning (i.e. not all parameters are treated as questionable). NSBL employs the concept of automatic relevance determination (ARD) to facilitate sparsity among questionable parameters through parameterized prior distributions. The analytical tractability of NSBL is enabled by employing Gaussian ARD priors and by building a Gaussian mixture-model approximation of the posterior parameter distribution that excludes the contribution of ARD priors. Subsequently, type-II maximum likelihood is executed using Newton's method whereby the evidence and its gradient and Hessian information are computed in a semi-analytical fashion. We show numerically and analytically that SBL is a special case of NSBL for linear regression models. Subsequently, a linear regression example involving multimodality in both parameter posterior pdf and model evidence is considered to demonstrate the performance of NSBL in cases where SBL is inapplicable. Next, NSBL is applied to identify sparsity among the damping coefficients of a mass-spring-damper model of a shear building frame. These numerical studies demonstrate the robustness and efficiency of NSBL in alleviating overfitting during Bayesian inversion of nonlinear physics-based models.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2014

Soil carbon sensitivity to temperature and carbon use efficiency compared across microbial-ecosystem models of varying complexity

Global ecosystem models may require microbial components to accurately predict feedbacks between climate warming and soil decomposition, but it is unclear what parameters and levels of complexity are ideal for scaling up to the globe. Here we conducted a model comparison using a conventional model with first-order decay and three microbial models of increasing complexity that simulate short- to long-term soil carbon dynamics. We focused on soil carbon responses to microbial carbon use efficiency (CUE) and temperature. Three scenarios were implemented in all models: constant CUE (held at 0.31), varied CUE ( 0.016 C 1), and 50 % acclimated CUE ( 0.008 C 1). Whereas the conventional model always showed soil carbon losses with increasing temperature, the microbial models each predicted a temperature threshold above which warming led to soil carbon gain. The location of this threshold depended on CUE scenario, with higher temperature thresholds under the acclimated and constant scenarios. This result suggests that the temperature sensitivity of CUE and the structure of the soil carbon model together regulate the long-term soil carbon response to warming. Equilibrium soil carbon stocks predicted by the microbial models were much less sensitive to changing inputs compared to the conventional model. Although many soil carbon dynamics were similar across microbial models, the most complex model showed less pronounced oscillations. Thus, adding model complexity (i.e. including enzyme pools) could improve the mechanistic representation of soil carbon dynamics during the transient phase in certain ecosystems. This study suggests that model structure and CUE parameterization should be carefully evaluated when scaling up microbial models to ecosystems and the globe.

Li, Jianwei↗