SEARCH · Engineering Papers

Results for “multivariate analysis”

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 757 records · Page 42

All sources NASA NTRS DOE OSTI

NASA NTRS1994

Extended cooperative control synthesis

This paper reports on research for extending the Cooperative Control Synthesis methodology to include a more accurate modeling of the pilot's controller dynamics. Cooperative Control Synthesis (CCS) is a methodology that addresses the problem of how to design control laws for piloted, high-order, multivariate systems and/or non-conventional dynamic configurations in the absence of flying qualities specifications. This is accomplished by emphasizing the parallel structure inherent in any pilot-controlled, augmented vehicle. The original CCS methodology is extended to include the Modified Optimal Control Model (MOCM), which is based upon the optimal control model of the human operator developed by Kleinman, Baron, and Levison in 1970. This model provides a modeling of the pilot's compensation dynamics that is more accurate than the simplified pilot dynamic representation currently in the CCS methodology. Inclusion of the MOCM into the CCS also enables the modeling of pilot-observation perception thresholds and pilot-observation attention allocation affects. This Extended Cooperative Control Synthesis (ECCS) allows for the direct calculation of pilot and system open- and closed-loop transfer functions in pole/zero form and is readily implemented in current software capable of analysis and design for dynamic systems. Example results based upon synthesizing an augmentation control law for an acceleration command system in a compensatory tracking task using the ECCS are compared with a similar synthesis performed by using the original CCS methodology. The ECCS is shown to provide augmentation control laws that yield more favorable, predicted closed-loop flying qualities and tracking performance than those synthesized using the original CCS methodology.

Davidson, John B.↗

NASA NTRS2013

An Alternative Flight Software Trigger Paradigm: Applying Multivariate Logistic Regression to Sense Trigger Conditions Using Inaccurate or Scarce Information

In late 2014, NASA will fly the Orion capsule on a Delta IV-Heavy rocket for the Exploration Flight Test-1 (EFT-1) mission. For EFT-1, the Orion capsule will be flying with a new GPS receiver and new navigation software. Given the experimental nature of the flight, the flight software must be robust to the loss of GPS measurements. Once the high-speed entry is complete, the drogue parachutes must be deployed within the proper conditions to stabilize the vehicle prior to deploying the main parachutes. When GPS is available in nominal operations, the vehicle will deploy the drogue parachutes based on an altitude trigger. However, when GPS is unavailable, the navigated altitude errors become excessively large, driving the need for a backup barometric altimeter to improve altitude knowledge. In order to increase overall robustness, the vehicle also has an alternate method of triggering the parachute deployment sequence based on planet-relative velocity if both the GPS and the barometric altimeter fail. However, this backup trigger results in large altitude errors relative to the targeted altitude. Motivated by this challenge, this paper demonstrates how logistic regression may be employed to semi-automatically generate robust triggers based on statistical analysis. Logistic regression is used as a ground processor pre-flight to develop a statistical classifier. The classifier would then be implemented in flight software and executed in real-time. This technique offers improved performance even in the face of highly inaccurate measurements. Although the logistic regression-based trigger approach will not be implemented within EFT-1 flight software, the methodology can be carried forward for future missions and vehicles.

Smith, Kelly M.↗

NASA NTRS2013

An Alternative Flight Software Paradigm: Applying Multivariate Logistic Regression to Sense Trigger Conditions using Inaccurate or Scarce Information

Smith, Kelly↗

DOE OSTI2025

Multivariate environmental and trait-based controls of transpiration in the Central Amazon Rainforest

Tropical forest tree mortality is increasing due to more severe droughts, yet our understanding of how tree traits and life strategies are linked to drought stress has been limited by measurement scarcity. The BIONTE (BIOmass and NuTrient Experiment) near Manaus, Brazil hosts one of the world’s largest sap flow installations, with sensors in 90 canopy trees across a wood density gradient monitored since June 2022. The 2023 El Niño drought provided a unique opportunity to evaluate how water availability impacts tree transpiration. An interpretable machine learning framework was used to study the complex interactions between transpiration and multiple environmental variables such as soil water availability and vapor pressure deficit (VPD), and how these interactions vary with wood density and individual trees. We found varying responses of transpiration from different trees during the El Niño drought. Transpiration generally increased with temperature, with stronger effects in wetter areas and in trees with low to medium wood density. However, this response was modulated by stomatal sensitivity to VPD, which constrained transpiration under high atmospheric demand, particularly in intermediate-moisture area. The inflection in transpiration rate at high temperatures (>32°C) underscores the role of stomatal and hydraulic regulation in limiting water loss and protecting trees from excessive evaporative demand. Analysis of soil water contribution to transpiration revealed unimodal patterns in wetter area, with peak contributions near 0.45 cm 3 cm -3 of surface soil water and declining or flat responses beyond that threshold, suggesting a shift from water- to energy-limited transpiration. In contrast, drier areas exhibited limited transpiration sensitivity to soil water conditions and minimal trait-based variation in VPD responses, indicating supply-limited conditions. Despite higher wood density trees being generally more resilient, this study shows diverse tree drought resilience, prompting further investigation into the specific traits and dynamics between environmental variables in regulating transpiration and other physiological processes in trees.

Drought↗

NASA NTRS1972

Application of multivariable search techniques to structural design optimization

Multivariable optimization techniques are applied to a particular class of minimum weight structural design problems: the design of an axially loaded, pressurized, stiffened cylinder. Minimum weight designs are obtained by a variety of search algorithms: first- and second-order, elemental perturbation, and randomized techniques. An exterior penalty function approach to constrained minimization is employed. Some comparisons are made with solutions obtained by an interior penalty function procedure. In general, it would appear that an interior penalty function approach may not be as well suited to the class of design problems considered as the exterior penalty function approach. It is also shown that a combination of search algorithms will tend to arrive at an extremal design in a more reliable manner than a single algorithm. The effect of incorporating realistic geometrical constraints on stiffener cross-sections is investigated. A limited comparison is made between minimum weight cylinders designed on the basis of a linear stability analysis and cylinders designed on the basis of empirical buckling data. Finally, a technique for locating more than one extremal is demonstrated.

Jones, R. T.↗

DOE OSTI2021

EPIsembleVis: A geo-visual analysis and comparison of the prediction ensembles of multiple COVID-19 models

In this work, we present EPIsembleVis, a web-based comparative visual analysis tool for evaluating the consistency of multiple COVID-19 prediction models. Our approach analyzes a collection of COVID-19 predictions from different epidemiological models as an ensemble and utilizes two metrics to quantify model performance. These metrics include (a) prediction uncertainty (represented as the dispersion of predictions in each ensemble) and (b) prediction error (calculated by comparing individual model predictions with the recorded data). Through an interactive visual interface, our approach provides a data-driven workflow for (a) selecting and constructing the COVID-19 model prediction ensemble based on the spatiotemporal overlap of available predictions of multiple epidemiological models, (b) quantifying the model performance using both the uncertainty of each model prediction ensemble, and the error of each ensemble member that represents individual model predictions, and (c) visualizing the spatiotemporal variability in the projection performance of individual models using a suite of novel ensemble visualization techniques, such as the data availability map, a spatiotemporal textured-tile calendar, multivariate rose chart, and time-series leaflet glyph. We demonstrate the capability of our ensemble visual interface through a case study that investigates the performance of weekly COVID-19 predictions, which are provided through the COVID-19 Forecast Hub UMass-Amherst Influenza Forecasting Center of Excellence [47] for the United States and United States Territories. The EPIsembleVis tool is implemented using open-source web technologies and adaptive system design, rendering it interoperable with Elasticsearch and Kibana for automatically ingesting COVID-19 predictions from online repositories, and it is generalizable for analyzing worldwide projections from more epidemiological models.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2022

Direct uranium enrichment assay in gaseous uranium hexafluoride with laser induced breakdown spectroscopy

A set of eleven U atomic emission lines, between 421.099 and 421.460 nm, were characterized for direct enrichment assay of gaseous UF 6 samples with laser induced breakdown spectroscopy (LIBS). Several iterative, multivariate nonlinear spectral-fitting algorithms were evaluated for their efficacies to extract the enrichment information from the measured LIBS spectra. Furthermore, wavelength-dependent weight factors, which take into consideration of the spectral-line position in the fitting model, the determined isotopic ratio, and the isotopic shift of the emission line, are essential for the spectral-fitting model to function adequately. The analytical accuracies and precision were typically within 0.5% in absolute [ 235 U/( 235 U + 238 U)] ratios.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

DOE OSTI2020

Machine learning approaches for structural and thermodynamic properties of a Lennard-Jones fluid

Predicting the functional properties of many molecular systems relies on understanding how atomistic interactions give rise to macroscale observables. However, current attempts to develop predictive models for the structural and thermodynamic properties of condensed-phase systems often rely on extensive parameter fitting to empirically selected functional forms whose effectiveness is limited to a narrow range of physical conditions. Here, we illustrate how these traditional fitting paradigms can be superseded using machine learning. Specifically, we use the results of molecular dynamics simulations to train machine learning protocols that are able to produce the radial distribution function, pressure, and internal energy of a Lennard-Jones fluid with increased accuracy in comparison to previous theoretical methods. The radial distribution function is determined using a variant of the segmented linear regression with the multivariate function decomposition approach developed by Craven et al. [J. Phys. Chem. Lett. 11, 4372 (2020)]. The pressure and internal energy are determined using expressions containing the learned radial distribution function and also a kernel ridge regression process that is trained directly on thermodynamic properties measured in simulation. The presented results suggest that the structural and thermodynamic properties of fluids may be determined more accurately through machine learning than through human-guided functional forms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS2006

Skin Temperature Analysis and Bias Correction in a Coupled Land-Atmosphere Data Assimilation System

In an initial investigation, remotely sensed surface temperature is assimilated into a coupled atmosphere/land global data assimilation system, with explicit accounting for biases in the model state. In this scheme, an incremental bias correction term is introduced in the model's surface energy budget. In its simplest form, the algorithm estimates and corrects a constant time mean bias for each gridpoint; additional benefits are attained with a refined version of the algorithm which allows for a correction of the mean diurnal cycle. The method is validated against the assimilated observations, as well as independent near-surface air temperature observations. In many regions, not accounting for the diurnal cycle of bias caused degradation of the diurnal amplitude of background model air temperature. Energy fluxes collected through the Coordinated Enhanced Observing Period (CEOP) are used to more closely inspect the surface energy budget. In general, sensible heat flux is improved with the surface temperature assimilation, and two stations show a reduction of bias by as much as 30 Wm(sup -2) Rondonia station in Amazonia, the Bowen ratio changes direction in an improvement related to the temperature assimilation. However, at many stations the monthly latent heat flux bias is slightly increased. These results show the impact of univariate assimilation of surface temperature observations on the surface energy budget, and suggest the need for multivariate land data assimilation. The results also show the need for independent validation data, especially flux stations in varied climate regimes.

Bosilovich, Michael G.↗

NASA NTRS2000

Ground Vibration Test Planning and Pre-Test Analysis for the X-33 Vehicle

This paper describes the results of the modal test planning and the pre-test analysis for the X-33 vehicle. The pre-test analysis included the selection of the target modes, selection of the sensor and shaker locations and the development of an accurate Test Analysis Model (TAM). For target mode selection, four techniques were considered, one based on the Modal Cost technique, one based on Balanced Singular Value technique, a technique known as the Root Sum Squared (RSS) method, and a Modal Kinetic Energy (MKE) approach. For selecting sensor locations, four techniques were also considered; one based on the Weighted Average Kinetic Energy (WAKE), one based on Guyan Reduction (GR), one emphasizing engineering judgment, and one based on an optimum sensor selection technique using Genetic Algorithm (GA) search technique combined with a criteria based on Hankel Singular Values (HSV's). For selecting shaker locations, four techniques were also considered; one based on the Weighted Average Driving Point Residue (WADPR), one based on engineering judgment and accessibility considerations, a frequency response method, and an optimum shaker location selection based on a GA search technique combined with a criteria based on HSV's. To evaluate the effectiveness of the proposed sensor and shaker locations for exciting the target modes, extensive numerical simulations were performed. Multivariate Mode Indicator Function (MMIF) was used to evaluate the effectiveness of each sensor & shaker set with respect to modal parameter identification. Several TAM reduction techniques were considered including, Guyan, IRS, Modal, and Hybrid. Based on a pre-test cross-orthogonality checks using various reduction techniques, a Hybrid TAM reduction technique was selected and was used for all three vehicle fuel level configurations.

Bedrossian, Herand↗

NASA NTRS1989

A methodology for formulating a minimal uncertainty model for robust control system design and analysis

In the design and analysis of robust control systems for uncertain plants, the technique of formulating what is termed an M-delta model has become widely accepted and applied in the robust control literature. The M represents the transfer function matrix M(s) of the nominal system, and delta represents an uncertainty matrix acting on M(s). The uncertainty can arise from various sources, such as structured uncertainty from parameter variations or multiple unstructured uncertainties from unmodeled dynamics and other neglected phenomena. In general, delta is a block diagonal matrix, and for real parameter variations the diagonal elements are real. As stated in the literature, this structure can always be formed for any linear interconnection of inputs, outputs, transfer functions, parameter variations, and perturbations. However, very little of the literature addresses methods for obtaining this structure, and none of this literature addresses a general methodology for obtaining a minimal M-delta model for a wide class of uncertainty. Since have a delta matrix of minimum order would improve the efficiency of structured singular value (or multivariable stability margin) computations, a method of obtaining a minimal M-delta model would be useful. A generalized method of obtaining a minimal M-delta structure for systems with real parameter variations is given.

Belcastro, Christine M.↗

DOE OSTI2023

Limited potential of irrigation to prevent potato yield losses in Germany under climate change

Climate models project higher temperatures and a substantial net water deficit during the summer months over Germany until the end of the 21st century. The agricultural sector is particularly vulnerable to changing weather and climate conditions. In this study, we examine the role of irrigation for potato yields as a climate change adaptation measure in Northeast Lower Saxony (Germany). The region represents the largest irrigated area in Germany and is one of the main growing areas of potatoes. We develop a statistical multivariate regression model to investigate whether regional climate and irrigation data at the monthly and county level are suitable to describe potato yield variability from 1978 to 2018. Irrigation is estimated based on the climatic water balance and crop water demand. These estimates are calibrated against irrigation data from local agencies to account for realistic monthly water withdrawals. We analyze future yield changes based on climate model projections for the 21st century and different irrigation scenarios. We find that the potato yield will decrease significantly by 18% on average until 2050 with climate projection uncertainty of ±6.4% under the high emission scenario RCP 8.5 and without irrigation. Irrigation at current levels could compensate for about 54% of these yield losses. Doubling the amount of irrigation minimizes yield losses by as much as 80%. Under the low (middle) emission scenario RCP 2.6 (RCP 4.5), potato yield are projected to decrease moderately by 4.5% ± 3.4% (5.4% ± 2.8%) with irrigation at current levels. We show that the yield losses mainly originate from higher temperatures during the summer. Our analysis indicates that the capacity of irrigation to minimize yield losses is limited. Strengthening climate change mitigation efforts may be more effective in preventing yield losses. However, the effect of elevated CO 2 might buffer the high yield losses, especially in RCP 8.5. The studies supports farmers and decision makers in assessing future risks in potato yields. It can guide policymakers in allocating economic resources to cope with climate change.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2022

Long‐Term Observations of Turbulence Vertical Velocity Spectra in a Convective Mixed Layer: Dependence on Land‐Surface Forcing in the U.S. Southern Great Plains

Doppler lidar vertical velocity retrievals were analyzed for the scale and structure of mixed-layer turbulence over a 7-year period, on fair-weather warm season days (442 cases) in the U.S. Southern Great Plains. Based on the hypothesized influence of land-surface forcing and updraft size on convective boundary layer clouds, a spectral analysis was performed to quantify effects of surface forcing (surface buoyancy flux, friction velocity, and evaporative fraction) on turbulence scale. Significant (order-of-magnitude) variations in spectral density were found in the energy-production subrange and mesoscale regimes. Integral scale decreased with increasing buoyancy flux (and Monin-Obukhov stability parameter), while spectral density in the energy-production subrange increased, implying a transition to buoyancy-driven cellular structures with narrower updrafts. The influence of stability parameter was limited to the neutral to convective transition, and could not explain the wide variation in spectral density in the mesoscale regime. However, high friction velocity (u* > 0.5 m s –1 ) was associated with larger integral scale (updraft width), and a greater portion of spectral density in the mesoscale regime. Lidar profiles and radar reflectivity for individual cases revealed evidence of roll and wave-like structures contributing to mesoscale variability on high friction velocity days, suggesting shear instabilities on days with wetter land surface conditions and smaller buoyancy flux. The role of friction velocity emphasizes the multivariate nature of surface influences on turbulence, and implies that variance-based turbulence closures may not be adequate for capturing effects of surface forcing on updraft size.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2001

Mapping Arid Vegetation Species Distributions in the White Mountains, Eastern California, Using AVIRIS, Topography, and Geology

Our challenge is to model plant species distributions in complex montane environments using disparate sources of data, including topography, geology, and hyperspectral data. From an ecologist's point of view, species distributions are determined by local environment and disturbance history, while spectral data are 'ancillary.' However, a remote sensor's perspective says that spectral data provide picture of what vegetation is there, topographic and geologic data are ancillary. In order to bridge the gap, all available data should be used to get the best possible prediction of species distributions using complex multivariate techniques implemented on a GIS. Vegetation reflects local climatic and nutrient conditions, both of which can be modeled, allowing predictive mapping of vegetation distributions. Geologic substrate strongly affects chemical, thermal, and physical properties of soils, while climatic conditions are determined by local topography. As elevation increases, precipitation increases and temperature decreases. Aspect, slope, and surrounding topography determine potential insolation, so that south-facing slopes are warmer and north-facing slopes cooler at a given elevation. Topographic position (ridge, slope, canyon, or meadow) and slope angle affect sediment accumulation and soil depth. These factors combine as complex environmental gradients, and underlie many features of plant distributions. Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) data, digital elevation models, digitized geologic maps, and 378 ground control points were used to predictively map species distributions in the central and southern White Mountains, along the western boundary of the Basin and Range province. Minimum Noise Fraction (MNF) bands were calculated from the visible and near-infrared AVIRIS bands, and combined with digitized geologic maps and topographic variables using Canonical Correspondence Analysis (CCA). CCA allows for modeling species 'envelopes' in multidimensional environmental space, which can then be projected across entire landscapes.

VandeVen, C.↗

DOE OSTI2024

Market optimization and technoeconomic analysis of hydrogen-electricity coproduction systems

Decarbonization efforts across North America, Europe, and beyond rely on variable renewable energy sources such as wind and solar, as well as alternative fuels, such as hydrogen, to support the sustainable energy transition. These advancements have prompted a need for more flexibility in the electric grid to complement non-dispatchable energy sources and increased demand from electrification. Integrated energy systems are well suited to provide this flexibility, but conventional technoeconomic modeling paradigms neglect the time-varying dynamic nature of the grid and thus undervalue resource flexibility. In this work, we develop a computational optimization framework for dynamic market-based technoeconomic comparison of integrated energy systems that coproduce low-carbon electricity and hydrogen (e.g., solid oxide fuel cells, solid oxide electrolysis) against technologies that only produce electricity (e.g., natural gas combined cycle with carbon capture) or only produce hydrogen. Our framework starts with rigorous physics-based process models, built in the open-source Institute for the Design of Advanced Energy Systems (IDAES) modeling and optimization platform, for six energy process concepts. Using these rigorous models and a workflow to optimally design each technology, the framework is shown to be capable of evaluating new and emerging technologies in varying energy markets under a plethora of future scenarios (i.e., renewables penetration, carbon tax, etc.). Ultimately, our framework finds that solid oxide fuel cell-based coproduction systems achieve positive profits for 85% of the analyzed market scenarios. From these market optimization results, we use multivariate linear regression (R 2 values up to 0.99) to determine which electricity price statistics are most significant to predict the optimized annual profit of each system. The proposed framework provides a powerful tool for directly comparing flexible, multi-product energy process concepts to help discern optimal technology and integration options.

08 HYDROGEN↗

NASA NTRS1993

Modeling of topographic effects on Antarctic sea ice using multivariate adaptive regression splines

The role of seafloor topography in the spatial variations of the southern ocean sea ice cover as observed (every other day) by the Nimbus 7 scanning multichannel microwave radiometer satellite in the years 1980, 1983, and 1984 is studied. Bottom bathymetry can affect sea ice surface characteristics because of the basically barotropic circulation of the ocean south of the Antarctic Circumpolar current. The main statistical tool used to quantify this effect is a local nonparametric regression model of sea ice concentration as a function of the depth and its first two derivatives in both meridional and zonal directions. First, we model the relationship of bathymetry to sea ice concentration in two sudy areas, one over the Maud Rise and the other over the Ross Sea shelf region. The multiple correlation coefficient is found to average 44% in the Maud Rise study area and 62% in the Ross Sea study area over the years 1980, 1983, and 1984. Second, a strategy of dividing the entire Antarctic region into an overlapping mosaic of small areas, or windows is considered. Keeping the windows small reduces the correlation of bathymetry with other factors such as wind, sea temperature, and distance to the continent. We find that although the form of the model varies from window to window due to the changing role of other relevant environmental variables, we are left with a spatially consistent ordering of the relative importance of the topographic predictors. For a set of three representative days in the Austral winter of 1980, the analysis shows that an average of 54% of the spatial variation in sea ice concentration over the entire ice cover can be attributed to topographic variables. The results thus support the hypothesis that there is a sea ice to bottom bathymetry link. However this should not undermine the considerable influence of wind, current, and temperature which affect the ice distribution directly and are partly responsible for the observed bathymetric effects.

De Veaux, Richard D.↗

DOE OSTI2024

Metal–support interactions in metal oxide-supported atomic, cluster, and nanoparticle catalysis

Supported metal catalysts are essential to a plethora of processes in the chemical industry. The overall performance of these catalysts depends strongly on the interaction of adsorbates at the atomic level, which can be manipulated and controlled by the different constituents of the active material (i.e., support and active metal). The description of catalyst activity and the relationship between active constituent and the support, or metal–support interactions (MSI), in heterogeneous (thermo)catalysts is a complex phenomenon with multivariate (dependent and independent) contributions that are difficult to disentangle, both experimentally and theoretically. So-called “strong metal–support interactions” have been reported for several decades and summarized in excellent review articles. However, in recent years, there has been a proliferation of new findings related to atomically dispersed metal sites, metal oxide defects, and, for example, the generation and evolution of MSI under reaction conditions, which has led to the designation of (sub)classifications of MSI deserving to be critically and systematically evaluated. These include dynamic restructuring under alternating redox and reaction conditions, adsorbate-induced MSI, and evidence of strong interactions in oxide-supported metal oxide catalysts. Here, we review recent literature on MSI in oxide-supported metal particles to provide an up-to-date understanding of the underlying physicochemical principles that dominate the observed effects in supported metal atomic, cluster, and nanoparticle catalysts. Critical evaluation of different subclassifications of MSI is provided, along with discussions on the formation mechanisms, theoretical and characterization advances, and tuning strategies to manipulate catalytic reaction performance. We also provide a perspective on the future of the field, and we discuss the analysis of different MSI effects on catalysis quantitatively.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Evaluating proxies for the drivers of natural gas productivity using machine-learning models

We report the extensive development of unconventional reservoirs using horizontal drilling and multistage hydraulic fracturing has generated large volumes of reservoir characterization and production data. The analysis of this abundant data using statistical methods and advanced machine-learning (ML) techniques can provide data-driven insights into well performance. Most predictive modeling studies have focused on the impact that different well completion and stimulation strategies have on well production but have not fully exploited the available in situ rock property data to determine its role in reservoir productivity. We have used machine-learning techniques to rank rock mechanical properties, microseismic attributes, and stimulation parameters in the order of their significance for predicting natural gas production from an unconventional reservoir. The data for this study came from a hydraulically fractured well in the Marcellus Shale in Monongalia County, West Virginia. The data classes included measurements aggregated by well completion stage that included (1) gas production, (2) well-log-derived measurements including bulk density, elastic moduli, shear impedance, compressional impedance, brittleness, and gamma measurements, (3) microseismic attributes, (4) long-period long-duration (LPLD) event counts, (5) fracture counts, and (6) stimulation parameters that included the fluid injection volume and average pumping pressure. To identify observable proxies for the drivers of gas production, we evaluated five commonly used ML approaches including multivariate adaptive regression spline, Gaussian mixture model, random forest, gradient boosting, and neural network. We selected five variables including LPLD event count, seismogenic b-value, hydraulic diffusivity, cumulative moment, and fluid volume as the features most likely to impact gas productivity at the stage level in the study area. The data-driven selection of these parameters for their importance in determining gas production can help reservoir engineers design more effective hydraulic-fracture treatments in the Marcellus Shale and other similar unconventional reservoirs. Plain language summary: We use machine-learning methods and data-driven selection of reservoir parameters to rank and better understand their importance in determining gas production, which can help reservoir engineers design more effective hydraulic-fracture treatments in the Marcellus Shale and other similar unconventional reservoirs.

58 GEOSCIENCES↗