Search NASA⌕ Search

SEARCH · Search NASA

Results for “model performance”

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 559 records · Page 31
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Deep-learning-based canopy height model generation from sub-meter resolution panchromatic satellite imagery

Canopy height models (CHMs) with sufficient resolution to distinguish individual trees are useful for a variety of applications. However, standard techniques to acquire such data, such as airborne lidar surveying, are often prohibitively expensive. Deep learning techniques for generating CHMs from high-resolution imagery are an attractive option to reduce costs. To date, success with these methods has been demonstrated using multichannel aerial photography and specialized satellite data products derived from multiple sensors, neither of which is commonly available at temporal resolutions finer than one year. Here we demonstrate a method to generate sub-meter resolution CHMs in three forests in California using a more abundant data source: sub-meter resolution, panchromatic satellite imagery from a single sensor. We show that phenology and species composition play important roles in model transferability; when trained using imagery from a single conifer forest in autumn, the model performs well on autumn imagery from a second conifer forest several hundred kilometers distant with no re-training. With modest additions to the training dataset, the same model generates minimally biased estimates of canopy height in both conifer and deciduous forests during multiple seasons. Because the model operates on satellite data with global coverage and a relatively short return interval, we propose its suitability to extrapolate tree-level canopy height data to remote regions and conduct high-temporal resolution monitoring of forest structure. We furthermore demonstrate the workflow’s applicability to fire modeling by conducting simulations in forests populated by trees measured using both this approach and airborne lidar surveying. We find minimal differences in fire behavior relative to a baseline case in which only statistical distributions of tree height and crown area are known. This result underscores the value of forest structural information derived from our workflow for improving the fidelity of wildland fire simulations, among other ecological applications.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2026

Data for Yield from Iowa’s first commercial miscanthus fields: implications of spatial variability for productivity and sustainability beyond research plots

This dataset contains biomass yield measurements and associated vegetation index data collected from commercial Miscanthus × giganteus fields in eastern Iowa during the 2022–2023 growing seasons. The data support the analyses presented in the article: “Yield From Iowa's First Commercial Miscanthus Fields: Implications of Spatial Variability for Productivity and Sustainability Beyond Research Plots.” We collected 105 ground-truth biomass samples from four mature commercial fields (>4 years old) covering 92.81 ha. Samples were taken from 3 m² quadrats that were hand-harvested in alignment with commercial harvest timing. Stem biomass (excluding leaves) was weighed, moisture-corrected, and converted to dry-matter yield expressed in Mg DM ha⁻¹. Sampling locations were selected to capture spatial variability visible in aerial imagery and were recorded using RTK GPS. Each biomass observation was paired with vegetation indices derived from high-resolution PlanetScope satellite imagery (3 m resolution). Images were acquired throughout the growing season, and indices were calculated to evaluate their ability to predict end-of-season biomass yield. Statistical and machine learning approaches were used to identify key predictors, and a linear regression model based on end-of-July Green Normalized Difference Vegetation Index (GNDVI) was developed and evaluated. This repository includes the data used in that modeling workflow. Management practices, economic data, full imagery time series, and additional methodological details are described in the associated publication and are not included here. The dataset consists of three comma-separated value (CSV) files: 1. Combine_Groundtruth_Yield_VI_22_23.csv This file contains ground-truth biomass yield measurements and associated key vegetation index values collected during the 2022 and 2023 growing seasons. Rows: 105 observations Columns: Year — Year of observation (2022 or 2023) Field — Field location identifier Sample_number — Unique sample identifier GNDVI_End_Jul — Green Normalized Difference Vegetation Index calculated at end of July GNDVI_End_Aug — Green Normalized Difference Vegetation Index calculated at end of August NDRE_End_Aug — Normalized Difference Red Edge index calculated at end of August Biomass_Stem_Yield_MgDM/ha — Measured stem biomass yield (megagrams dry matter per hectare) 2. trainData_GNDVI.csv This file contains the subset of observations used to train the predictive relationship between July GNDVI and biomass yield. Rows: 76 observations Columns: Unnamed: 0 — Row index retained from the original data processing workflow GNDVI_End_Jul — GNDVI at end of July Stem_Yield_MgDM/ha — Observed stem biomass yield (Mg DM ha⁻¹) 3. testData_GNDVI.csv This file contains the test dataset used to evaluate model performance. Rows: 29 observations Columns: Unnamed: 0 — Row index retained from the original data processing workflow GNDVI_End_Jul — GNDVI at end of July Predicted_Yield_MgDM/ha — Model-predicted stem biomass yield (Mg DM ha⁻¹) Observed_Yield_MgDM/ha — Measured stem biomass yield (Mg DM ha⁻¹)

Potential yield, yield gap, in-field management, y↗
DOE OSTI2024

Investigating Material Properties of Subsurface Rock Formations Modified by Engineering Mineral Precipitation (Final Scientific and Technical Report)

Montana State University’s (MSU) Energy Research Institute (ERI), in collaboration with the Center for Biofilm Engineering (CBE) and the Department of Civil Engineering (CE), has conducted a long‐term research program aimed at developing a novel cementing agent to address wellbore integrity and reduce the unwanted upward migration of fluids and greenhouse gases from the subsurface. The primary technology developed through this research program is known as ureolysis‐induced calcite precipitation (UICP), which harnesses bio‐chemical processes to precipitate calcium carbonate (CaCO 3 ). The same general process can also be called microbially-induced calcium carbonate precipitation (MICP) when microbes provide the process-catalyzing urease enzyme. Both terms are used in this report. Results have conclusively demonstrated that, if properly controlled, UICP can successfully seal fractures, high permeability zones, and compromised cement in the vicinity of wellbores and in nearby caprock. This technology has been successfully deployed to mitigate annular leakage in two test wells and over sixty commercial wells with a 100% success rate. This success in downhole deployment generates consideration of other subsurface applications where UICP could provide benefit to the energy sector, such as shale property modification for unconventional oil and gas recovery. The focus of this research project was to investigate fundamental material and mechanical properties of select shale cores and analyze how these properties change due to engineered mineral precipitation with the intent to control these properties to achieve a range of engineering objectives. Ultimately, the project aim was to identify valuable new areas where application of UICP might contribute to national energy security and environmental protection. The research workplan coupled UICP treatment of core samples, nuclear magnetic resonance (NMR) characterization, and mechanical strength testing at MSU with advanced X‐Ray micro-computed tomography (μCT) imaging and numerical modeling performed by collaborators at two national laboratories, the National Energy Technology Laboratory (NETL) and Lawrence Berkeley National Laboratory (LBNL). Experimental results are useful to inform geo-mechanical models which could be applied to predict mineralized rock formation behavior at field scale. Our findings suggest that NMR and μCT methods to detect and quantify biomineral formation in shale fractures are complementary and consistent with each other. Either could be used to estimate the volume of new mineral formed by UICP in shale fractures. The use of surfactants and guar gum to enhance biomineral precipitation in shale fractures merits further research. UICP can, under some conditions, increase the tensile strength of sealed shale fractures beyond that of the intact shale. These findings demonstrate that continued research in this area may be valuable to understanding and improving shale resource recovery techniques.

58 GEOSCIENCES↗
DOE OSTI2022

Using Maximum Circular Velocity in Halo Occupation Distribution Models to Predict Galaxy Clustering

Abstract Standard halo occupation distribution (HOD) models predict how galaxies occupy halos based on a single property, the virial mass, M vir . To incorporate galaxy assembly bias, we consider alternative halo mass proxies to use in the HOD. A promising replacement is maximum circular velocity, V max , which measures the halo’s potential well, and successfully predicts galaxy clustering in subhalo abundance matching models. We fit galaxy clustering measurements from SDSS DR7 with two modified HODs besides the standard M vir -based model: one where V max is used as the halo property that determines galaxy occupation, and the other adopting an adjustable combination of M vir and V max . Except for the reduced volume M r < − 20 threshold sample fit with V max alone, neither change improves the model performance in predicting the projected two-point correlation function, w p . We conclude that switching the primary variable to V max does not significantly improve the HOD model’s ability to fit galaxy clustering.

Mezini, Lorena↗
DOE OSTI2023

Battery state-of-health diagnostics during fast cycling using physics-informed deep-learning

Rapid, in-situ Li-ion battery state-of-health (SOH) quantification is challenging. Li-ion battery aging can vary significantly with chemistry, operating conditions, cycling demands, electrode design, and operation history. As a cell ages, optimal and safe operating conditions need to be adapted to account for battery degradation by tracking critical aging modes such as loss-of-lithium-inventory (LLI), loss-of-active-material (LAM) in either electrode, and/or impedance rise. This manuscript describes a framework for identifying battery aging modes in-operando using fast-rate voltage charge/discharge responses. The framework uses a physically based Li-ion battery model to produce synthetic high-rate responses at aged states. The aging model is calibrated against experimental data from cells with different electrode loadings and cycled under a variety of fast-charging conditions (1 h, 15 min, 10 min, and 7 min charging). The synthetically generated high-rate responses at aged states are then used to train a deep-learning model to identify real cell state-of-health from fast charge/discharge battery voltage responses. The synthetically trained deep-learning model performance is validated by comparing to standard incremental capacity analysis and half-cell measurements. Finally, the framework demonstrates the benefits of using high-rate physics-based models to generate synthetic data for training deep-learning models.

25 ENERGY STORAGE↗
DOE OSTI2024

Ensemble methods for quantification of potassium oxide in ChemCam Mars and laboratory spectra

In this paper we test new approaches for predicting the amount of element oxides in rock samples from the ChemCam instrument suite onboard the NASA Curiosity rover by focusing on K 2 O. Using the expanded dataset compiled by Gasda et al. (2021) with and without the Earth to Mars (E2M and NoE2M) transformation discussed in Clegg et al. (2017) we trained blended submodels using the “double blending” technique and compared these to ensemble methods (Random Forest, ExtraTrees, and Gradient Boosting Regression). We found that ensemble methods performed similar to blended submodels when looking at RMSE-P on the laboratory spectra and provided significant advantages when looking at spectra coming from Mars. For the full model, blended submodels achieved an RMSE-P of 0.62 and 0.60 (E2M and NoE2M respectively) while Gradient Boosting Regression resulted in a slightly improved RMSE-P of 0.59 and 0.60. More importantly, by employing a local RMSE-P estimation technique where model performance is evaluated based on nearby test samples we found that using ensemble methods can lower the quantification limit for K 2 O from the current value of ≈0.6 wt% to ≈0.08 wt% using Extra Trees and Random Forest. This would allow for a much larger range of K 2 O values to be quantified on Mars with greater certainty given that most targets seen on Mars tend to have <1 wt% K2O. Finally, we used both Mean Decrease in Impurity (MDI) and permutation importance techniques to investigate the wavelengths used by the ensemble methods and found that they correspond to known potassium emission lines. This suggests that ensemble methods can provide an easier to train and improved alternative to blended submodels for predicting potassium compositions from Laser Induced Breakdown Spectroscopy (LIBS) data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Highly Integrated WBG Power Module for Next Generation Plug-In Vehicles

Current inverter technologies used in electric vehicles (EVs) are bulky, heavy, and expensive. The performance of the inverter is also limited by the traditional Si-based semiconductor and packaging technology. These constraints are responsible for extra weight and volume in the vehicle that reduces the driving range along with usable space within the vehicle. Moreover, operating temperature restrictions, and boundaries on the inverter location within the vehicle drastically reduces the inverter performance. Thus an improved inverter design is required for next generation electric vehicles. The objective of this project was to research, develop and demonstrate a highly integrated WBG power module targeting the next generation of plugin hybrid electric vehicles (PHEVs). This power module needed to be functionally and mechanically optimized for General Motors (GM) traction inverter architecture. General Motors was the prime contractor, led the project, and oversaw the design, build, and test of the integrated power module. Virginia Polytechnic Institute and State University (VT) performed modeling and simulation tasks as well as power module design support. Monolith Semiconductor, Inc provided SiC dies and consulted on devices for the inverter. Oak Ridge National Laboratory (ORNL) performed power module technology evaluations, design evaluations of interconnect options, and performed material analysis and cross-sectioning of die activities.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2024

Segmentation and Classification of Fission as Pores in Reactor Irradiated Annular U–10Zr Metallic Fuel Using Machine Learning Models

Metallic fuels, particularly U—10Zr, are promising candidates for next-generation sodium-cooled fast reactors. Irradiation of nuclear fuels in reactors can lead to the formation of solid and gas fission product which subsequently forms microstructural pores, deteriorating fuel performance. Due to the massive amount of pores and complex phases formed, a quantitative description of fission gas pores is not yet available, preventing the development of microstructure-informed fuel performance modeling for fuel qualification. This paper applied a pre-trained deep learning model to ~10,260 high magnification scanning electron microscopy images. This method increased the accuracy of fission gas pore segmentation and allows statistical features to be extracted which cannot be achieved manually. A pre-trained decision tree model worked on the segemenation results and further classified the pores into different categories to produce a correlation between the pores, movement of lanthanides, and temperature gradient during irradiation. Finally, this paper emphasizes the potentials of machine learning models to accelerate fuel research, development, and qualification for advanced reactors.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Digital polycrystalline microstructure generation using diffusion probabilistic models

Accurate micromechanical simulation of polycrystalline materials requires a realistic digital representation of the grain scale microstructure. Here, this work demonstrates the use of a generative diffusion probabilistic model for synthesizing single phase polycrystalline realizations. The model performs well and is capable of producing realistic microstructures consisting of not just simple equiaxed structures but also structures exhibiting more complex spatial arrangements. Masked microstructure generation reveals that the model is context aware of morphological descriptors which may be encoded in the latent space. Training on more diverse data sets, with scaled up architectures, may enable development of future models capable of synthesizing even more complex microstructural features.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Computational insights into hydrogen adsorption energies on medium-entropy oxides

High entropy oxides (HEOs) have emerged as promising catalysts for several important chemical transformations including alkane activation. Hydrogen adsorption energy (HAE) has been used as a key descriptor for many reactions including methane C–H activation and hydrogen evolution reactions. Hence, understanding the relationship between HAEs and the surface chemistry of HEO surfaces could lay the foundation for meaningful correlations among methane C–H activation, HAE, and the complex, local environment of HEO surfaces. Here, we used a medium-entropy oxide as a prototypical system – Mg 0.25 Ni 0.25 Cu 0.25 Zn 0.25 O with a rock-salt structure – to interrogate these relationships. We sampled 2000 different surfaces of its (100) plane and calculated the HAEs at randomly chosen surface O sites using density functional theory (DFT). Our analysis of the 2000 data points reveals that the HAEs at the surface O sites are significantly influenced by the local environment around the adsorption sites, particularly the nature of the metal atom directly below the surface O site where H adsorbs. After comparing several popular graph-neural-network-based machine learning models, we found that the DimeNet++ model performed best achieving satisfactory accuracy in predicting HAEs for both Mg 0.25 Ni 0.25 Cu 0.25 Zn 0.25 O and slightly varied compositions. Our work underscores the promise of such models and the need for further refinement to address the complexity of HEOs.

Song, Haohong [Vanderbilt Univ., Nashville, TN (Un↗
DOE OSTI2024

Phasor-Domain Modeling of Inverter Based Resources Considering Unbalanced Grid Faults

A significant increase in inverter-based distributed energy resources (DERs) is expected in distribution systems of the near future. The latest phasor modeling approach of DERs with emerging control technologies can capture relevant dynamics in large-scale three-phase distribution systems with unbalanced construction. However, there is a lack of appropriate phasor models to deal with unbalanced fault scenarios. As such, this paper describes per-phase phasor modeling approach appropriate for unbalanced grid disturbance studies in distribution systems with high penetration of DERs. The overall modeling performance of representative DERs is simulated in phasor domain GridLAB-D environment and validated against controller hardware-in-the loop (CHIL) testing or electromagnetic transient (EMT) based switching model simulation in PSCAD with test cases including different unbalanced grid faults. The high accuracy retained from the simulation verifies the effectiveness of the proposed phasor modeling approach considering interactions between positive and negative sequence signals.

Chen, Zhe↗
DOE OSTI2025

Heat Transfer in Void Generating Foam Decomposition: Further Development

Continued development of the additive conductivity material model, used to simulate changes in heat transfer that occurs in void generating foam decomposition, has resulted in an improved model and new features. The previous version of the model was calibrated against the Aria Bulk Fluid Element (BFE) solution and proposed a third-order polynomial correction term best captured the increased heat transfer due to voids in the foam. An investigation of the Fuego Conjugate Heat Transfer (CHT) and Aria BFE solutions at several geometries revealed the CHT solution and BFE solution had differing behavior across length scales, especially at smaller scales. Five calibration studies, using the Fuego CHT as the calibration data, were carried out with polynomial functions of 4-th, 3-rd, 2-nd, 1-st and 0-th orders to determine the best correction function that generalized well across length scales. Each polynomial function was calibrated/trained on six different sized geometries and then tested on three uniquely sized geometries. This study revealed that the 1-st order additive conductivity model performed the best. A new feature of void formation scaling was implemented to more realistically capture the heat transfer as voids are created. A scaling term was added to the model to activate the conductivity correction as decomposition progresses.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Quantifying Leaf Chlorophyll Concentration of Sorghum from Hyperspectral Data Using Derivative Calculus and Machine Learning

Leaf chlorophyll concentration (LCC) is an important indicator of plant health, vigor, physiological status, productivity, and nutrient deficiencies. Hyperspectral spectroscopy at leaf level has been widely used to estimate LCC accurately and non-destructively. This study utilized leaf-level hyperspectral data with derivative calculus and machine learning to estimate LCC of sorghum. We calculated fractional derivative (FD) orders starting from 0.2 to 2.0 with 0.2 order increments. Additionally, 43 common vegetation indices (VIs) were calculated from leaf spectral reflectance factor to make comparisons with reflectance-based data. Within the modeling pipeline, three feature selection methods were assessed: Pearson’s correlation coefficient (PCC), partial least squares based variable importance in the projection (VIP), and random forest-based mean decrease impurity (MDI). Finally, we used partial least squares regression (PLSR), random forest regression (RFR), support vector regression (SVR), and extreme learning regression (ELR) to estimate the LCC of sorghum. Results showed that: (1) increasing derivative order can show improved model performance until certain order for reflectance-based analysis; however, it is inconclusive to state that a particular order is optimal for estimating LCC of sorghum; (2) VI-based modeling outperformed derivative augmented reflectance factor-based modeling; (3) mean decrease impurity was found effective in selecting sensitive features from large feature space (reflectance-based analysis), whereas simple Pearson’s correlation coefficient worked better with smaller feature space (VI-based analysis); and (4) SVR outperformed all other models within reflectance-based analysis; alternatively, ELR with VIs from original reflectance yielded slightly better results compared to all other models.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

Toolpath Optimization of Polymer Composite Additive Manufacturing AM Printed Mold to Improve Performance for Compression Molding Applications

This study explores the optimization of toolpaths for short fiber polymer composite additive manufacturing (AM) printed molds in compression molding applications, an emerging trend driven by advancements in 3D printing technology. AM printed molds offer reduced tooling costs and greater design flexibility, particularly for prototyping and low-volume production. The proposed optimization approach evaluates mold performance using two key metrics: deformation and shape accuracy. Multiple toolpaths are generated, and finite element (FE) simulations capture fiber orientation and material properties, which are then integrated into a mechanical performance model. Shape accuracy is assessed by comparing the deformed mold shape with the original using surface normals as a descriptor. The analysis identifies the toolpath that delivers the best overall performance, establishing it as the most suitable option for achieving optimal mold characteristics in compression molding applications.

Limaye, Madhura [ORNL]↗
DOE OSTI2025

Moltensaltpropnet

MoltenSaltPropnet is a physics-informed machine learning framework that aims to predict the thermophysical properties of molten fluoride and chloride salt mixtures, which are crucial for the design and safety of Generation IV molten salt reactors. The code processes data from the Molten-Salt Thermal Properties Database (MSTDB-TP) and the Janz compendium, converting critically evaluated correlations into fast, differentiable surrogate models for density, viscosity, thermal conductivity, and heat capacity across 448 distinct salt systems. The implementation consists of several key components: 1. Data Curation: The code parses and cleans the raw data, normalizing elemental mole fractions and extracting relevant regression coefficients for various thermophysical properties. 2. Feature Engineering: It generates fixed-length numerical descriptors that encapsulate the composition and temperature, incorporating polynomial interaction terms and dimensionality-reduction techniques to optimize model performance. 3. Coefficient Learning: Four different machine learning architectures are employed: a deep residual network (ResNet), a Kolmogorov–Arnold network (KAN), a sparsity-inducing neural network (SNN), and classical regression models. Each model learns to predict coefficients that define the temperature-dependent correlations for the thermophysical properties. 4. Property Reconstruction: The predicted coefficients are used to compute temperature-dependent property values, ensuring positivity and monotonic trends through a composite loss function that enforces physical constraints. 5. User Interface: An open-source web application enables users to filter the database, train task-specific models, and visualize the results, allowing for rapid exploration of candidate salt mixtures. MoltenSaltPropnet bridges the gap between limited experimental data and high-fidelity reactor simulations, providing a powerful tool for researchers in the field of molten salt reactors and advanced nuclear energy systems.

Retamales, Mauricio Eduardo Tano [Idaho National L↗
DOE OSTI2023

Quantum Neural Networks: Issues, Training, and Applications

Our work in the field aims at explaining the limitations and expressive power of Quantum Machine Learning models, as well as finding feasible training algorithms that could be implemented in near-term Quantum Computers. The promise of Quantum Machine Learning is that by incorporating quantum effects, such as entanglement, into machine learning models researchers can improve model performance and understand more complex datasets. This pledge is particularly pronounced in the design of Quantum neural networks (QNNs), a promising framework for creating quantum algorithms, that promise to outperform classical models by combining the speedups of quantum computation with the widespread successes of deep learning. We show that applying this approach alone to quantum deep learning is problematic given that an excess of entanglement between the hidden and visible layers can destroy the predictive power of our QNN models. We address the barren plateau problem by suggesting the use of a generative, unbounded, nonlinear loss function with simple gradients. The loss function quantifies how much the quantum states generated by the QNNs differ from the data and the goal during training is to minimize it. Finally, we showcase how to use generative training to construct a "classical-quantum" neural network to accurately interpolate between the ground states of a Molecular Hamiltonian, a central question in Quantum Chemistry.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Green AI: Insights Into Deep Learning's Looming Energy Efficiency Crisis

As demands grow to integrate artificial intelligence into every aspect of industry, commerce, and life, deep learning's exploding energy cost has become a looming crisis, making AI systems a salient energy-efficiency challenge. One might expect that doubling a neural network's size would halve its error rate, or at least allow it to achieve greater performance given the same amount of time and energy. I will present clear and substantial scientific evidence which indicates that not only is this intuition wildly wrong, but that neural networks scale so poorly that to increase deep learning performance by only a small fraction can easily require an order of magnitude or more increase in computational resources and energy. Further, the marginal trade-off price of to increase model performance rapidly explodes as performance targets are increased. To address this challenge, I will provide a toolkit of techniques that can be applied today to mitigate the inefficiency of modern deep learning. And, I will conclude by illuminating a practical path forward towards efficient, Green AI.

artificial intelligence↗
DOE OSTI2021

Predicting measures of soil health using the microbiome and supervised machine learning

Soil health encompasses a range of biological, chemical, and physical soil properties that sustain the commercial and ecological value of agroecosystems. Monitoring soil health requires a comprehensive set of diagnostics that can be cost-prohibitive for routine analyses. The soil microbiome provides a rich source of information about soil properties, which can be assayed in a high-throughput, cost-effective way. We evaluated the accuracy of random forest (RF) and support vector machine (SVM) regression and classification models in predicting 12 measures of soil health, tillage status, and soil texture from 16S rRNA gene amplicon data with an operationally relevant sample set. We validated the efficacy of the best performing models against independent datasets and also tested best practices for processing microbiome data for use in machine learning. Soil health metrics could be predicted from microbiome data with the best models achieving a Kappa value of ~0.65, for categorical assessments, and a R2 value of ~0.8, for numerical scores. Biological health ratings were better predicted than chemical or physical ratings. Validation with independent datasets revealed that models had general predictive value for soil properties, including yield. The ecological profiles of several taxa important for model accuracy matched the observed relationships with soil health, including Pyrinomonadaceae, Nitrososphaeraceae, and Candidatus Udeaobacter. Models trained at the highest taxonomic resolution proved most accurate, with losses in accuracy resulting from rarefying, sparsity filtering, and aggregating at higher taxonomic ranks. Furthermore, our study provides the groundwork for developing scalable technology to use microbiome-based diagnostics for the assessment of soil health.

16S rRNA gene↗