Search NASA⌕ Search

SEARCH · Search NASA

Results for “evaluation datasets”

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 721 records · Page 40

Trait-Based Representation of Biological Nitrification: Model Development, Testing, and Predicted Community Composition

Trait-based microbial models show clear promise as tools to represent the diversity and activity of microorganisms across ecosystem gradients. These models parameterize specific traits that determine the relative fitness of an “organism” in a given environment, and represent the complexity of biological systems across temporal and spatial scales. In this study we introduce a microbial community trait-based modeling framework (MicroTrait) focused on nitrification (MicroTrait-N) that represents the ammonia-oxidizing bacteria (AOB) and ammonia-oxidizing archaea (AOA) and nitrite-oxidizing bacteria (NOB) using traits related to enzyme kinetics and physiological properties.We used this model to predict nitrifier diversity, ammonia (NH3) oxidation rates, and nitrous oxide (N2O) production across pH, temperature, and substrate gradients. Predicted nitrifier diversity was predominantly determined by temperature and substrate availability, the latter was strongly influenced by pH. The model predicted that transient N2O production rates are maximized by a decoupling of the AOB and NOB communities, resulting in an accumulation and detoxification of nitrite to N2O by AOB. However, cumulative N2O production (over 6 month simulations) is maximized in a system where the relationship between AOB and NOB is maintained.When the reactions uncouple, the AOB become unstable and biomass declines rapidly, resulting in decreased NH3 oxidation and N2O production. We evaluated this model against site level chemical datasets from the interior of Alaska and accurately simulated NH3 oxidation rates and the relative ratio of AOA:AOB biomass. The predicted community structure and activity indicate (a) parameterization of a small number of traits may be sufficient to broadly characterize nitrifying community structure and (b) changing decadal trends in climate and edaphic conditions could impact nitrification rates in ways that are not captured by extant biogeochemical models.

59 BASIC BIOLOGICAL SCIENCES↗

Impact of light-absorbing particles on snow albedo darkening and associated radiative forcing over high-mountain Asia: high-resolution WRF-Chem modeling and new satellite observations

Light-absorbing particles (LAPs), mainly dust and black carbon, can significantly impact snowmelt and regional water availability over high-mountain Asia (HMA). In this study, for the first time, online aerosol–snow interactions are enabled and a fully coupled chemistry Weather Research and Forecasting (WRF-Chem) regional model is used to simulate LAP-induced radiative forcing on snow surfaces in HMA at relatively high spatial resolution (12 km, WRF-HR) compared with previous studies. Simulated macro- and microphysical properties of the snowpack and LAP-induced snow darkening are evaluated against new spatially and temporally complete datasets of snow-covered area, grain size, and impurity-induced albedo reduction over HMA. A WRF-Chem quasi-global simulation with the same configuration as WRF-HR buta coarser spatial resolution (1°, WRF-CR) is also used to illustrate the impact of spatial resolution on simulations of snow properties and aerosol distribution over HMA. Due to a more realistic representation of terrain slopes over HMA, the higher-resolution model (WRF-HR) shows significantly better performance in simulating snow area cover, duration of snow cover, snow albedo and snow grain size over HMA, as well as an evidently better atmospheric aerosol loading and mean LAP concentration in snow. However, the differences in albedo reduction from model and satellite retrievals is large during winter due to associated overestimation in simulated snow fraction. It is noteworthy that Himalayan snow cover has high magnitudes of LAP-induced snow albedo reduction (4% - 8%) in pre-monsoon seasons (both from WRF-HR and satellite estimates), which induces a snow-mediated radiative forcing of ~30-50 W m -2 . As a result, the Himalayas (specifically the western Himalayas) hold the most vulnerable glaciers and mountain snowpack to the LAP-induced snow darkening effect within HMA. In summary, coarse spatial resolution and absence of snow–aerosol interactions over the Himalayan cryosphere will result insignificant underestimation of aerosol effects on snow melting and regional hydroclimate.

54 ENVIRONMENTAL SCIENCES↗

An ensemble of 48 physically perturbed model estimates of the 1/8° terrestrial water budget over the conterminous United States, 1980–2015

Terrestrial water budget (TWB) data over large domains are of high interest for various hydrological applications. Spatiotemporally continuous and physically consistent estimations of TWB rely on land surface models (LSMs). As an augmentation of the operational North American Land Data Assimilation System Phase 2 (NLDAS-2) four-LSM ensemble, this paper describes a dataset simulated from an ensemble of 48 physics configurations of the Noah LSM with multi-physics options (Noah-MP). The 48 Noah-MP physics configurations are selected to give a representative cross-section of commonly used LSMs for parameterizing runoff, atmospheric surface layer turbulence, soil moisture limitation on photosynthesis, and stomatal conductance. The dataset spans from 1980 to 2015 over the conterminous United States (CONUS) at a monthly temporal resolution and a 1/8° spatial resolution. The dataset variables include total evapotranspiration and its constituents (canopy evaporation, soil evaporation, and transpiration), runoff (the surface and subsurface components), as well as terrestrial water storage (snow water equivalent, four-layer soil water content from the surface down to 2 m, and the groundwater storage anomaly). The dataset is available at https://doi.org/10.5281/zenodo.7109816. Evaluations carried out in this study and previous investigations show that the ensemble per forms well in reproducing the observed terrestrial water storage, snow water equivalent, soil moisture, and runoff. Noah-MP complements the NLDAS models well, and adding Noah-MP consistently improves the NLDAS es timations of the above variables in most areas of CONUS. Besides, the perturbed-physics ensemble facilitates the identification of model deficiencies. The parameterizations of shallow snow, spatially varying groundwater dynamics, and near-surface atmospheric turbulence should be improved in future model versions.

54 ENVIRONMENTAL SCIENCES↗

Tropospheric Ozone Over a Tropical Atlantic Station in the Northern Hemisphere: Paramaribo, Surinam (6 deg N, 55 deg W)

We present an analysis of 2.5 years of weekly ozone soundings conducted at a new monitoring station in Paramaribo, Surinam (6 deg N,55 deg W). This is currently one of only three ozone sounding stations in the northern hemisphere (NH) tropics, and the only one in the equatorial Atlantic region. Paramaribo is part of the Southern Hemisphere ADditional Ozone Sounding program (SHADOZ). Due to its position close to the equator, the Inter Tropical Convergence Zone (ITCZ) passes over Paramaribo twice per year, which results in a semi-annual seasonality of many parameters including relative humidity and ozone. The dataset from Paramaribo is used to: (1) evaluate ozone variability relative to precipitation, atmospheric circulation patterns and biomass burning; (2) contrast ozone at the NH equatorial Atlantic with that at nearby southern hemisphere (SH) stations Natal (6 deg S,35 deg W) and Ascension (8 deg S,14 deg W); (3) compare the seasonality of tropospheric ozone with a satellite-derived ozone product: Tropical Tropospheric Ozone Columns from the Modified Residual method (MR-TTOC). We find that Paramaribo is a distinctly Atlantic station. Despite its position north of the equator, it resembles nearby SH stations during most of the year. Transport patterns in the lower and middle troposphere during February and March differ from SH stations, which leads to a seasonality of ozone with two maxima. MR-TTOC over Paramaribo does not match the observed seasonality of ozone due to the use of a SH ozone sonde climatology in the MR method. The Paramaribo ozone record is used to suggest an improvement for northern hemisphere MR-TTOC retrievals. We conclude that station Paramaribo shows unique features in the region, and clearly adds new information to the existing SHADOZ record.

Peters, W.↗

Accuracy and Precision in the Southern Hemisphere Additional Ozonesondes (SHADOZ) Dataset 1998-2000 in Light of the JOSIE-2000 Results

A network of 12 southern hemisphere tropical and subtropical stations in the Southern Hemisphere ADditional OZonesondes (SHADOZ) project has provided over 2000 profiles of stratospheric and tropospheric ozone since 1998. Balloon-borne electrochemical concentration cell (ECC) ozonesondes are used with standard radiosondes for pressure, temperature and relative humidity measurements. The archived data are available at:http: //croc.gsfc.nasa.gov/shadoz. In Thompson et al., accuracies and imprecisions in the SHADOZ 1998- 2000 dataset were examined using ground-based instruments and the TOMS total ozone measurement (version 7) as references. Small variations in ozonesonde technique introduced possible biases from station-to-station. SHADOZ total ozone column amounts are now compared to version 8 TOMS; discrepancies between the two datasets are reduced 2\% on average. An evaluation of ozone variations among the stations is made using the results of a series of chamber simulations of ozone launches (JOSIE-2000, Juelich Ozonesonde Intercomparison Experiment) in which a standard reference ozone instrument was employed with the various sonde techniques used in SHADOZ. A number of variations in SHADOZ ozone data are explained when differences in solution strength, data processing and instrument type (manufacturer) are taken into account.

Witte, J. C.↗

AstroLoc: An Efficient and Robust Localizer for a Free-flying Robot

We present AstroLoc, an efficient and robust monocular visual-inertial graph-based localization system used by the Astrobee free-flying robots onboard the International Space Station (ISS). We provide a novel localization system that limits the traditionally higher computation times for graph-based localization systems and enables the resource constrained Astrobee robots to benefit from their increased accuracy. We also introduce methods for handling cheirality issues for visual odometry and localization factors that further increase localization robustness. We evaluate the performance of AstroLoc on a dataset of ISS activities and show that it greatly improves pose, velocity, and IMU bias estimation accuracy while efficiently running in a limited computation environment. The source code for AstroLoc is released to the public.

Localization↗

Exercising Prediction Validation Methods for Analysis of Quiet Supersonic Overflight

En route noise certification procedures for quiet supersonic aircraft are in development under the International Civil Aviation Organization for a proposed standard that could enable civil supersonic overland flight. While these quiet supersonic aircraft do not exist yet, methods are being developed and evaluated using existing N-wave empirical datasets and low-boom simulation datasets. One potential aspect of future certification methods is the ability to use a validated prediction model to predict the certification noise level under reference day weather conditions. Methods to validate predictions, such as the two-sample t-test, Fisher’s combined probability test, two-sample Kolmogorov-Smirnov test, and equivalence testing, have been exercised and modified for this application. While some methods compare mean values, others compare values across the distribution, and a combination of methods can be used. The methods were applied to a subset of an N-wave dataset to demonstrate the applicability and identify potential areas for improvement. Atmospheric turbulence variability was included as a factor in these analyses through KZKFourier propagation simulations with parameters matching actual flights and weather conditions. Results are presented for six noise metrics (PL, ASEL, BSEL, DSEL, ESEL, and ISBAP), and plans to expand the analyses are discussed.

sonic boom↗

Mastering HPC Runtime Prediction: From Observing Patterns to a Methodological Approach: Preprint

The continual expansion of high-performance computing (HPC) brings with it an increasing need for efficiency. Heavy investment in energy, hardware, and software infrastructure to support peta- and exascale computing requires the optimization of existing systems and, wherever possible, the discernment and adoption of best-practices towards these goals. Such is the case for runtime prediction. When a job is submitted to an HPC system, an estimate of its runtime is provided by the user in the form of "requested wallclock''. Error in this user-provided estimate can lead to jobs being prematurely killed by the scheduler, increased wait time on the queue, and decreased system utilization. More than fifteen years of research has been directed at mitigating these effects by using data-driven runtime predictions. Codified here is a set of commonalities and insights emerging from this body of work, which we present as recommendations and best practices. These practices are combined into a methodological approach described and evaluated on an 11-million-job dataset from the National Renewable Energy Laboratory's petascale HPC system, Eagle. This dataset and the accompanying codebase have been released to the public domain for the benefit of the wider HPC research community.

high performance computing↗

Enhancing Operational Safety via Agentic Dialogue Hazard Identification Analysis

Operational safety in high-stakes domains such as industrial process control, autonomous, and safety-critical systems demand reliable hazard identification. While large language models (LLMs) have shown promise in automating safety analysis tasks, single-turn, monolithic inference is brittle: it lacks the self-correction, deliberation, and contextual refinement that safety engineers apply iteratively. In this paper, we introduce HAZDIAL, a framework that investigates whether structured agentic dialogue (multi-agent, multi-turn interactions) improves the quality of NLP-based hazard identification over single-pass baselines. We systematically compare two dialogue modalities: adversarial debate and constructive discussion, and propose an genetic algorithm-based agentic interaction optimization. We evaluate all configurations against a curated golden dataset using standard classification metrics (accuracy, precision, recall, F1) and a novel dialogue metrics. This work advances the intersection of dialogue systems, multi-agent reasoning, and AI safety, providing empirical evidence for dialogue-driven hazard analysis.

Das, Sanjay [ORNL] (ORCID:0009000542591915)↗

Dataset describing two reference models for full-spectral lighting and daylight simulations together with implementations for two software systems

A dataset of two spectral lighting simulation reference models - one office and one factory hall - is presented. It aims to demonstrate and support full-spectral daylight and electric lighting simulations and facilitate evaluation of non-visual effects of light. The dataset includes Rhino CAD geometry, comprehensive spectral material and light source data and window system BSDF data. Example implementations in the two software tools, Radiance and OWL, enable reproducible workflows and support adoption in other software. The dataset is openly available on Zenodo. The office model reproduces Room 518 at the University of Innsbruck, including a west-facing façade and interior furnishings. The factory hall model follows the proposed geometry in the European standard 15193 for building energy performance. Interior reflectances in the office were measured in-situ using a handheld spectrometer. Exterior spectra and factory hall materials matching specified reflectances were obtained from an online spectral materials database. Glazing transmittance was derived from IGDB data using LBNL Optics/WINDOW. BSDFs for venetian blinds at various tilt angles, and for a diffusing pane adapted from the Complex Glazing Database, were generated in WINDOW. Luminaires in both models are specified with photometric files (Eulumdat/IES) and lamp spectra (Fluorescent 840, 4000 K LED). The provided example implementations (Radiance, OWL) include prepared input data and scripts to run first spectral simulations; example results are also included. The dataset is prepared to support reuse by researchers, designers and software developers for method validation, software engineering and comparison, and development of spectral metrics and controls.

Geisler-Moroder, David↗

Visualizing Clustering and Uncertainty Analysis with Multivariate Longitudinal Data

Multivariate time-series datasets are intrinsic to the study of dynamic, naturalistic behavior, such as in the applications of finance and motion video analysis. Statistical models provide the ability to identify event patterns in these data under conditions of uncertainty, but researchers must be able to evaluate how well a model uses available information in a dataset for clustering decisions and for uncertainty information. The Hidden Markov Model (HMM) is an established method for clustering time-series data, where the hidden states of the HMM are the clusters. We develop novel methods for quantifying the uncertainty of the performance of and for visualizing the clustering performance and uncertainty of fitting a HMM to multivariate time-series data. We explain the usefulness of uncertainty quantification and visualization with evaluating the performance of clustering models, as well as how information exploitation of time-series datasets can be enhanced. We implement our methods to cluster patterns of scanpaths from raw eye tracking data.

97 MATHEMATICS AND COMPUTING↗

RIPL - Reference Input Parameter Library for Calculation of Nuclear Reactions and Nuclear Data Evaluations

We describe the physics and data included in the Reference Input Parameter Library, which is devoted to input parameters needed in calculations of nuclear reactions and nuclear data evaluations. Advanced modelling codes require substantial numerical input, therefore the International Atomic Energy Agency (IAEA) has worked extensively since 1993 on a library of validated nuclear-model input parameters, referred to as the Reference Input Parameter Library (RIPL). A final RIPL coordinated research project (RIPL-3) was brought to a successful conclusion in December 2008, after 15 years of challenging work carried out through three consecutive IAEA projects. The RIPL-3 library was released in January 2009, and is available on the Web through (http://www-nds.iaea.org/RIPL-3/). This work and the resulting database are extremely important to theoreticians involved in the development and use of nuclear reaction modelling (ALICE, EMPIRE, GNASH, UNF, TALYS) both for theoretical research and nuclear data evaluations. The numerical data and computer codes included in RIPL-3 are arranged in seven segments: MASSES contains ground-state properties of nuclei for about 9000 nuclei, including three theoretical predictions of masses and the evaluated experimental masses of Audi et al. (2003). DISCRETE LEVELS contains 117 datasets (one for each element) with all known level schemes, electromagnetic and {gamma}-ray decay probabilities available from ENSDF in October 2007. NEUTRON RESONANCES contains average resonance parameters prepared on the basis of the evaluations performed by Ignatyuk and Mughabghab. OPTICAL MODEL contains 495 sets of phenomenological optical model parameters defined in a wide energy range. When there are insufficient experimental data, the evaluator has to resort to either global parameterizations or microscopic approaches. Radial density distributions to be used as input for microscopic calculations are stored in the MASSES segment. LEVEL DENSITIES contains phenomenological parameterizations based on the modified Fermi gas and superfluid models and microscopic calculations which are based on a realistic microscopic single-particle level scheme. Partial level densities formulae are also recommended. All tabulated total level densities are consistent with both the recommended average neutron resonance parameters and discrete levels. GAMMA contains parameters that quantify giant resonances, experimental gamma-ray strength functions and methods for calculating gamma emission in statistical model codes. The experimental GDR parameters are represented by Lorentzian fits to the photo-absorption cross sections for 102 nuclides ranging from {sup 51}V to {sup 239}Pu. FISSION includes global prescriptions for fission barriers and nuclear level densities at fission saddle points based on microscopic HFB calculations constrained by experimental fission cross sections.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Continental Scale Hydrostratigraphy: Basin‐Scale Testing of Alternative Data‐Driven Approaches

Abstract Integrated hydrological modeling is an effective method for understanding interactions between parts of the hydrologic cycle, quantifying water resources, and furthering knowledge of hydrologic processes. However, these models are dependent on robust and accurate datasets that physically represent spatial characteristics as model inputs. This study evaluates multiple data‐driven approaches for estimating hydraulic conductivity and subsurface properties at the continental‐scale, constructed from existing subsurface dataset components. Each subsurface configuration represents upper (unconfined) hydrogeology, lower (confined) hydrogeology, and the presence of a vertical flow barrier. Configurations are tested in two large‐scale U.S. watersheds using an integrated model. Model results are compared to observed streamflow and steady state water table depth (WTD). We provide model results for a range of configurations and show that both WTD and surface water partitioning are important indicators of performance. We also show that geology data source, total subsurface depth, anisotropy, and inclusion of a vertical flow barrier are the most important considerations for subsurface configurations. While a range of configurations proved viable, we provide a recommended Selected National Configuration 1 km resolution subsurface dataset for use in distributed large‐and continental‐scale hydrologic modeling.

54 ENVIRONMENTAL SCIENCES↗

Time-Based CAN IDS Paper Results Code

Modern vehicles are complex cyber-physical systems made of hundreds of electronic control units (ECUs) that communicate over controller area networks (CANs). This inherited complexity has expanded the CAN attack surface which is vulnerable to message injection attacks. These injections change the overall timing characteristics of messages on the bus, and thus, to detect these malicious messages, time-based intrusion detection systems (IDSs) have been proposed. However, time-based IDSs are usually trained and tested on low-fidelity datasets with unrealistic, labeled attacks. This makes difficult the task of evaluating, comparing, and validating IDSs. Here we detail and benchmark four time-based IDSs against the newly published ROAD dataset, the first open CAN IDS dataset with real (non-simulated) stealthy attacks with physically verified effects. We found that methods that perform hypothesis testing by explicitly estimating message timing distributions have lower performance than methods that seek anomalies in a distribution related statistic. In particular, these “distribution-agnostic” based methods outperform “distribution-based” methods by at least 55% in area under the precision-recall curve (AUC-PR). Our results expand the body of knowledge of CAN time-based IDSs by providing details of these methods and reporting their results when tested on datasets with real advanced attacks. Finally, we develop an after-market plug-in detector using lightweight hardware, which can be used to deploy the best performing IDS method on nearly any vehicle.

Moriano, Pablo [Oak Ridge National Lab. (ORNL), Oa↗

Multi-frequency, multi-polarization SAR and radiometer sea ice classification

Microwave remotely sensed data were acquired simultaneously in March 1988 over Beaufort, Chukchi, and Bering Sea ice using the JPL multifrequency, polarimetric synthetic aperture radar (SAR), and multifrequency dual-polarization GSFC radiometer. The experiment was conducted as a series of coordinated underflights of the DMSP SSM/I satellite radiometer to validate ice products derived from its radiance values. Concurrent flights by an NRL P-3 aircraft enabled overlapping high-resolution, single-frequency images to be acquired over the same surfaces using a Ka-band scanning microwave radiometer. Precise coregistration to an accuracy of 100 m, +/- 25 m enables comparison of temporally and spatially coincident active and passive microwave datasets. Three-channel polarimetric SAR data are compared with dual-polarized radiometer data at several SSM/I frequencies, showing backscatter signatures to be highly correlated with radiometric temperatures. Further comparisons are used to evaluate ice concentration retrievals and to show how multiparameter datasets can be used to resolve ambiguities present in identifying certain sea ice forms with a single microwave instrument.

Drinkwater, Mark R.↗

pvOps: a Python package for empirical analysis of photovoltaic field data

The purpose of pvOps is to support empirical evaluations of data collected in the field related to the operations and maintenance (O&M) of photovoltaic (PV) power plants. pvOps presently contains modules that address the diversity of field data, including text-based maintenance logs, current-voltage (IV) curves, and timeseries of production information. The package functions leverage machine learning, visualization, and other techniques to enable cleaning, processing, and fusion of these datasets. These capabilities are intended to facilitate easier evaluation of field patterns and extraction of relevant insights to support reliability-related decision-making for PV sites. The open-source code, examples, and instructions for installing the package through PyPI can be accessed through the GitHub repository.

14 SOLAR ENERGY↗

Revisiting Five Years of CASMI Contests with EPA Identification Tools

Software applications for high resolution mass spectrometry (HRMS)-based non-targeted analysis (NTA) continue to enhance chemical identification capabilities. Given the variety of available applications, determining the most fit-for-purpose tools and workflows can be difficult. The Critical Assessment of Small Molecule Identification (CASMI) contests were initiated in 2012 to provide a means to evaluate compound identification tools on a standardized set of blinded tandem mass spectrometry (MS/MS) data. Five CASMI contests have resulted in recommendations, publications, and invaluable datasets for practitioners of HRMS-based screening studies. The US Environmental Protection Agency’s (EPA) CompTox Chemicals Dashboard is now recognized as a valuable resource for compound identification in NTA studies. However, this application was too new and immature in functionality to participate in the five previous CASMI contests. In this work, we performed compound identification on all five CASMI contest datasets using Dashboard tools and data in order to critically evaluate Dashboard performance relative to that of other applications. CASMI data was accessed via the CASMI webpage and processed for use in our spectral matching and identification workflow. Relative to applications used by former contest participants, our tools, data, and workflow performed well, placing more challenge compounds in the top five of ranked candidates than did the winners of three contest years and tying in a fourth. In addition, we conducted an in-depth review of the CASMI structure sets and made these reviewed sets available via the Dashboard. Our results suggest that Dashboard data and tools would enhance chemical identification capabilities for practitioners of HRMS-based NTA.

97 MATHEMATICS AND COMPUTING↗

Predicting future well performance for environmental remediation design using deep learning

Here in this study, we developed a deep learning (DL) framework with a multi-channel three-dimensional convolutional neural network (MC3D-CNN) to predict well performance and thereby assist future environmental remediation design. Such prediction of extraction well performance at designated locations is critical for configuring pump-and-treat (P&T) well network design and operation, setting reasonable target closure dates for overall remedying, and estimating remedy costs. The framework is developed with operational and monitoring data routinely collected during P&T remedy operations, including well extraction and injection rates as well as in situ contaminant concentrations. Traditionally, the collected data were rarely used for purposes other than assessing past well performance and the accuracy of the conceptual site model. However, recent advances in data-driven computational approaches enable better use of the large datasets to inform future well performance, enhance site characterization, and improve remediation planning. In this study, we established a DL framework to integrate transient three-dimensional contaminant plumes and multiple aquifer properties (e.g., hydraulic conductivity and hydrostratigraphic maps) to identify characteristic patterns controlling and representing extraction well mass recovery, aiming at providing future mass recovery estimates for existing wells and candidate wells at any proposed locations. We evaluated our framework by using a realistic synthetic dataset generated from a well-calibrated flow and transport model used in the 200 West Area of the U.S. Department of Energy’s Hanford Site in southeastern Washington state. The multi-channel feature in our framework allows integration of various types and temporal densities of training datasets for DL model development. Overall, we found that the trained DL model achieved an accuracy of over 90% in ranking extraction well performance in validation datasets, and over 80% in predicting high-performance-ranking well locations. This data-informed approach provides a flexible tool to support adaptive site management, streamline decision-making, and potentially reduce remediation time and costs. Our DL framework can be used as a filtering tool to improve the current P&T network optimization design by reducing the number of candidate well locations.

54 ENVIRONMENTAL SCIENCES↗