Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data consistent”

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 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
NASA NTRS1973

Study of spacecraft direct readout meteorological systems

Characteristics are defined of the next generation direct readout meteorological satellite system with particular application to Tiros N. Both space and ground systems are included. The recommended space system is composed of four geosynchronous satellites and two low altitude satellites in sun-synchronous orbit. The goesynchronous satellites transmit to direct readout ground stations via a shared S-band link, relayed FOFAX satellite cloud cover pictures (visible and infrared) and weather charts (WEFAX). Basic sensor data is transmitted to regional Data Utilization Stations via the same S-band link. Basic sensor data consists of 0.5 n.m. sub-point resolution data in the 0.55 - 0.7 micron spectral region, and 4.0 n.m. resolution data in the 10.5 - 12.6 micron spectral region. The two low altitude satellites in sun-synchronous orbit provide data to direct readout ground stations via a 137 MHz link, a 400 Mhz link, and an S-band link.

Bartlett, R.↗
NASA NTRS1981

Combining LANDSAT MSS, aerial photographs and ground measurements to estimate rangeland productivity

The production of a vegetation map of over 2.2 million acres with detail down to the plant community level, and the production of estimates of rangeland productivity (pounds of usable forage per acre for cattle) for a 500,000 acre subset of area with a design goal for accuracy and precision of + or - 20% at the 80% confidence level, are considered. The data consist of five groups: maps of area, LANDSAT data, digital terrain data, large scale aerial photography, and ground plots. An outline of the data acquisition and data reduction schemes are presented.

Gialdini, M. J.↗
NASA NTRS2015

65 Years of Reprocessed GLDAS Version 2.0 Data and Their Exploration Using the NASA GES DISC Giovanni

GLDAS-2.0 data have been reprocessed with updated Princeton meteorological forcing data within the Land Information System (LIS) Version 7, and temporal coverage have been extended to 1948-2012.Global Land Data Assimilation System Version 2 (GLDAS-2) has two components: GLDAS-2.0: entirely forced with the Princeton meteorological forcing data GLDAS-2.1: forced with atmospheric analysis and observation-based data after 2001In order to create more climatologically consistent data sets, NASA GSFC's Hydrological Sciences Laboratory (HSL) has recently reprocessed the GLDAS-2.0, by using updated Princeton meteorological forcing data within the LIS Version 7.GLDAS-2.0 data and data services are provided at NASA GES DISC Hydrology Data and Information Services Center (HDISC), in collaboration with HSL.

data assimilation↗
NASA NTRSDate not supplied

Silicon Partitioning Between Iron-Rich Metal and Silicate Melt: New Constraints From Aerodynamic Laser Levitation Furnace Experiments

The partitioning of light elements (e.g., Si, C, H, etc.) between silicate melts and iron metal alloys as a function of pressure and temperature is a critical constraint in defining the compositional evolution of planetary interiors. While there are a significant number of experimental studies in the literature that examine the various influences of oxygen fugacity, temperature, pressure, and composition on the metal-melt partitioning of silicon, it is difficult to disentangle the effects of each variable due to the nominally independent parameters (e.g., pressure, temperature, composition) being collinear. To better test for the independent effects of pressure, temperature, and melt composition on the partitioning of silicon between iron alloy and silicate melt, new high temperature (1800°–2500°C), rapid quench (~700°C/s) aerodynamic levitation laser-heating experiments in the Fe-Si-Mg-O (En65-Fa5-Si3-Fe27) system were conducted. This high temperature, low pressure method is ideal for these experiments because it eliminates the metal container as a sink for elements of interest to partition into, and it has a rapid quench rate that minimizes potential exsolution of silicon from the metal phase during quench cooling. Given the short run times of the levitation experiments (30–60 seconds), a time series at a lower temperature (1700°C) in a vertical gas-mixing muffle tube furnace was also conducted to assess equilibrium times of the bulk composition. All successful runs were imaged using back-scatter electron imaging and analyzed for phase composition using a JEOL 8530F field emission microprobe. These new data were added to a comprehensive and internally consistent data set created from published experiments, and subsequently modeled using parameters designed to minimize collinearity (e.g., optical basicity is used as a proxy for melt composition). High collinearity is found in the data set however between temperature and oxygen fugacity, precluding any model that contains both as nominally independent variables. The resulting model demonstrates a statistically significant effect of pressure, temperature, and melt composition on the partitioning of silicon in Fe-rich metal, which has implications for the formation of planetary cores in rocky planets and provides important constraints on potential bulk core compositions.

Metal↗
DOE OSTI2023

Status of DUNE Offline Computing

We summarize the status of Deep Underground Neutrino Experiment (DUNE) Offline Software and Computing program. We describe plans for the computing infrastructure needed to acquire, catalog, reconstruct, simulate and analyze the data from the DUNE experiment and its prototypes in pursuit of the experiment's physics goals of precision measurements of neutrino oscillation parameters, detection of astrophysical neutrinos, measurement of neutrino interaction properties and searches for physics beyond the Standard Model. In contrast to traditional HEP computational problems, DUNE's Liquid Argon Time Projection Chamber data consist of simple but very large (many GB) data objects which share many characteristics with astrophysical images. We have successfully reconstructed and simulated data from 4% prototype detector runs at CERN. The data volume from the full DUNE detector, when it starts commissioning late in this decade will present memory management challenges in conventional processing but significant opportunities to use advances in machine learning and pattern recognition as a frontier user of High Performance Computing facilities capable of massively parallel processing. Our goal is to develop infrastructure resources that are flexible and accessible enough to support creative software solutions as HEP computing evolves.

43 PARTICLE ACCELERATORS↗
DOE OSTI2026

COMPASS-FME Terrestrial Ecosystem Manipulation to Probe the Effects of Storm Treatments (TEMPEST) Experiment Level 2 Sensor Data v2-1

This is the version v2-1 Level 2 (L2) data release for COMPASS-FME environmental sensors located at our Terrestrial Ecosystem Manipulation to Probe the Effects of Storm Treatments (TEMPEST) experimental site. This manipulative, ecosystem-scale TEMPEST experiment addresses the potential for freshwater and estuarine-water disturbance events to alter tree function, species composition, and ecosystem processes in a deciduous coastal forest in MD, USA. The experiment uses a large-unit (2000 m2), un-replicated experimental design, with three 50 m × 40 m plots serving as control, freshwater, and estuarine-water treatments. Level 2 (L2) data consist of sensor observations from the COMPASS-FME synoptic sites, TEMPEST, and DELUGE. Compared to the L1 data, these are more consistent (always 15-minute timestamps for the entire year); better QA/QC’d (out of bounds, out of service, and extreme outlier values are removed); and more complete, with a gap-filled time series available alongside the main observations, and additional derived (calculated) variables. L2 data are intended to be rapidly and easily usable in analyses and simulations. However, algorithmic outlier identification always carries the risk of removing valid data, and Level 1 data may be more suitable for analyses that focus on variability or extreme events. This dataset includes: - An overall dataset README file that describes the current version, gives citation and contact information, etc. - Site- and year-specific folders, each holding variable-specific Parquet (a high performance, space efficient format; see https://parquet.apache.org) data files for each site and plot in that year. - Metadata files within each site-year folder provide full information on data units, expected ranges, contact information, detailed flood times, as well as a general description of the site. - Environmental sensor types that appear in the data files include weather (ClimaVUE50, CS, RM Young, and LI instruments in the graphs below); soil conditions (TEROS12); soil redox state (Redox); groundwater variables (AquaTROLL200 and AquaTROLL600); open water sondes (Exo); tree sap velocity (Sapflow); and system voltage and state (Datalogger). Data are reported every 15 minutes. Please see v2-1 TEMPEST L2 Sensor Package Quick Start.pdf for detailed information on data package structure, temporal coverage, and versioning. Data files are in Apache Parquet, a high performance, space efficient format for tabular data. These files can be read using R's `arrow` package (https://arrow.apache.org/docs/r/), with similar tools available in other languages. The TEMPEST flood events occurred on the following dates. They lasted for ~10 hours each day and delivered ~80,000 gallons to each plot; many data streams are available at 1 or 5 minute frequency during these periods. * Tests: Aug 25 (fresh plot) and Sep 9 (salt plot), 2021 * TEMPEST 1: June 22, 2022 * TEMPEST 2: June 6-7, 2023 * TEMPEST 3: June 11-13, 2024

EARTH SCIENCE > ATMOSPHERE > ATMOSPHERIC TEMPERATU↗
DOE OSTI2026

COMPASS-FME Synoptic Sites Level 2 Sensor Data v2-1

This is the version 2-1 Level 2 (L2) data release for COMPASS-FME environmental sensors located at our synoptic field sites. COMPASS-FME is studying sites in two distinct regions, the Chesapeake Bay and the Western Lake Erie Basin. We established the network at seven "synoptic" (observational) sites along the Chesapeake Bay and Lake Erie coastlines, collectively generating over three million observations per month, to track and comprehend environmental changes where land and water intersect. Additionally, the two regions provide an interesting contrast of saltwater and freshwater coasts that allow us to differentiate the impacts of inundation and coastal water chemistries in two nationally important coastal systems. Level 2 (L2) data consist of sensor observations from the COMPASS-FME synoptic sites, TEMPEST, and DELUGE. Compared to the L1 data, these are more consistent (always 15-minute timestamps for the entire year); better QA/QC’d (out of bounds, out of service, and extreme outlier values are removed); and more complete, with a gap-filled time series available alongside the main observations, and additional derived (calculated) variables. L2 data are intended to be rapidly and easily usable in analyses and simulations. However, algorithmic outlier identification always carries the risk of removing valid data, and Level 1 data may be more suitable for analyses that focus on variability or extreme events. This dataset includes: - An overall dataset README file that describes the current version, gives citation and contact information, etc. - Site- and year-specific folders, each holding variable-specific Parquet (a high performance, space efficient format; see https://parquet.apache.org) data files for each site and plot in that year. - Metadata files within each site-year folder provide full information on data units, expected ranges, contact information, detailed flood times, as well as a general description of the site. - Environmental sensor types that appear in the data files include weather (ClimaVUE50, CS, RM Young, and LI instruments in the graphs below); soil conditions (TEROS12); soil redox state (Redox); groundwater variables (AquaTROLL200 and AquaTROLL600); open water sondes (Exo); tree sap velocity (Sapflow); and system voltage and state (Datalogger). Data are reported every 15 minutes. Data files are in Apache Parquet, a high performance, space efficient format for tabular data. These files can be read using R's `arrow` package (https://arrow.apache.org/docs/r/), with similar tools available in other languages. Please see v2-1 L2 Sensor Package QStart.pdf for detailed information on data package structure, temporal coverage, and versioning.

EARTH SCIENCE > ATMOSPHERE > ATMOSPHERIC TEMPERATU↗
DOE OSTI2023

DOLCE: A Model-Based Probabilistic Diffusion Framework for Limited-Angle CT Reconstruction

Limited-Angle Computed Tomography (LACT) is a non-destructive evaluation technique used in a variety of applications ranging from security to medicine. The limited angle coverage in LACT is often a dominant source of severe artifacts in the reconstructed images, making it a challenging inverse problem. We present DOLCE, a new deep model-based framework for LACT that uses a conditional diffusion model as an image prior. Diffusion models are a recent class of deep generative models that are relatively easy to train due to their implementation as image denoisers. DOLCE can form high-quality images from severely under-sampled data by integrating data-consistency updates with the sampling updates of a diffusion model, which is conditioned on the transformed limited-angle data. We show through extensive experimentation on several challenging real LACT datasets that, the same pre-trained DOLCE model achieves the SOTA performance on drastically different types of images. Additionally, we show that, unlike standard LACT reconstruction methods, DOLCE naturally enables the quantification of the reconstruction uncertainty by generating multiple samples consistent with the measured data.

Kim, Hyojin↗
NASA NTRS1982

Processing MAGSAT data for comparison with geoid anomalies

A digital data library of MAGSAT data consisting of 1,615,636 measurements from the quiet data set, is geographically sorted, and allows rapid analysis and processing of all the quiet magnetic data about any selected location. Because this library of MAGSAT data is compatible with existing gravity and geoid data library processing and display system software, correlations between MAGSAT, surface gravity, GEOS-3 radar altimeter geoid and bathymetric data sets can be conveniently detected and analyzed. Polynomial trends from each half-orbit were removed as an effective way of estimating and removing ring current effects following estimation of the core field contribution. It was found that a third order polynomial is the lowest polynomial order that appears to provide the best consistency of residual anomalies between coincident orbits.

Bowin, C. O.↗
NASA NTRSDate not supplied

Using Satellite Soil Moisture and Rainfall in the Landslide Hazard Assessment for Situational Awareness System

The Landslide Hazard Assessment for Situational Awareness system(LHASA)gives a global view of landslide hazard in nearly real time. Currently, it is being upgraded from version 1 to version 2, which entails improvements along several dimensions. These include the incorporation of new predictors, machine learning, and new event-based landslide inventories. As a result, LHASA version 2 substantially improves on the prior performanceand introduces a probabilistic element to the global landslide nowcast. Data from the soil moisture active-passive (SMAP) satellite has been assimilated into a globally consistent data product with a latency less than 3 days, known as SMAP Level 4. In LHASA, thesedata representthe antecedent conditions prior to landslide-triggering rainfall. In some cases, soil moisture may have accumulated over aperiod of many months. The model behind SMAP Level 4 also estimates the amount of snow on the ground, which is an important factor in some landslide events. LHASA also incorporates this information as an antecedent condition that modulates the response torainfall. Slope, lithology, and active faults were also used as predictor variables. These factors can have a strong influence on where landslides initiate.LHASA relies on precipitation estimates from the Global Precipitation Measurement mission to identify the locations where landslides are most probable. The low latency and consistent global coverage of these data make them ideal for real-time applications at continental to global scales. LHASA relies primarily on rainfall from the last 24 hours to spothazardous sites, which is rescaled by the local 99thpercentile rainfall.However, the multi-day latency of SMAP requires the use of a 2-day antecedent rainfall variable to represent the accumulation of rain between the antecedent soil moisture and current rainfall. LHASA merges these predictors with XGBoost, a commonly used machine-learning tool, relying on historical landslide inventories to develop the relationship between landslide occurrence and various risk factors. The resulting model relies heavily on current daily rainfall, but other factors also play an important role. LHASA outputsthe probability oflandslide occurrence ona grid of roughly one kilometer over all continents from 60 North to 60 South latitude. Evaluation over the period 2019-2020 showsthat LHASA version 2 doubles the accuracy of the global landslide nowcast without increasing the global false alarm rate. LHASA also identifies the areas where the human exposure to landslide hazard is most intense. Landslide hazard is divided into 4 levels: minimal, low, moderate, and high. Next, the number of persons and the length of major roads (primary and secondary roads)within each of these areas is calculated for every second-level administrative district (county). These results can be viewedthrough a web portal hosted at the Goddard Space Flight Center. In addition, users can download daily hazard and exposure data.LHASAversion 2uses machine learning and satellite data to identify areas of probable landslide hazard within hours of heavy rainfall. Itsglobal maps are significantly more accurate, and it now includes rapid estimates of exposed populations and infrastructure. In addition, a forecast mode will be implemented soon.

Thomas Stanley↗
NASA NTRS2007

Science Archives in the 21st Century: A NASA LAMBDA Report

Lambda is a thematic data center that focuses on serving the cosmic microwave background (CMB) research community. LAMBDA is an active archive for NASA's Cosmic Background Explorer (COBE) and Wilkinson Microwave Anisotropy Probe (WMAP) mission data sets. In addition, LAMBDA provides analysis software, on-line tools, relevant ancillary data and important web links. LAMBDA also tries to preserve the most important ground-based and suborbital CMB data sets. CMB data is unlike other astrophysical data, consisting of intrinsically diffuse surface brightness photometry with a signal contrast of the order 1 part in 100,000 relative to the uniform background. Because of the extremely faint signal levels, the signal-to-noise ratio is relatively low and detailed instrument-specific knowledge of the data is essential. While the number of data sets being produced is not especially large, those data sets are becoming large and complex. That tendency will increase when the many polarization experiments currently being deployed begin producing data. The LAMBDA experience supports many aspects of the NASA data archive model developed informally over the last ten years-that small focused data centers are often more effective than larger more ambitious collections, for example; that data centers are usually best run by active scientists; that it can be particularly advantageous if those scientists are leaders in the use of the archived data sets; etc. LAMBDA has done some things so well that they might provide lessons for other archives. A lot of effort has been devoted to developing a simple and consistent interface to data sets, for example; and serving all the documentation required via simple 'more' pages and longer explanatory supplements. Many of the problems faced by LAMBDA will also not surprise anyone trying to manage other space science data. These range from persuading mission scientists to provide their data as quickly as possible, to dealing with a high volume of nuisance (spam) messages. Because so many data center problems and solutions are common across individual data centers and disciplines it would be very valuable to establish some new systems of communication - such as informal email lists for administrators and developers. But resources are very limited, so new timeconsuming and inefficient mechanisms - like too-frequent and too-structured meetingsshould be avoided. Although there are great advantages to being small, agile and independent, there are also some areas where science data centers within and without NASA could be better coordinated - for the assignment of persistent identifiers; to encourage the early adoption of useful standards and technologies; etc. Some super-structure to facilitate such coordination might be beneficial as long as it doesn't begin to control the other work of the archives, and become a "methodology police". In this respect the CCSDS "Reference Model for an Open Archive Information System" is a little worrying. It may be that the closer a data center gets to following such a detailed prescription, the less effective it will become. It is much better to have an informal coordination process than a bureaucratic straight-jacket.

Butterworth, P.↗
NASA NTRS1977

A methodology for experimentally based determination of gap shrinkage and effective lifetimes in the emitter and base of p-n junction solar cells and other p-n junction devices

An experimentally based methodology that determines the effective gap shrinkage and lifetime in the emitter of a p-n junction solar cell is described which provides an experimental means for assessing the importance of gap shrinkage relative to that of large recombination rates in the highly doped emitter. The base lifetime is also determined. The methodology pertains to a solar cell after the junction is formed, so that each material parameter determined includes the effects of the processing used in junction fabrication. The methodology consists of strategy and procedures for designing experiments and interpreting data consistently with the physical mechanisms governing device behavior. This careful linking to the device physics uncover the material parameters concealed in the data. To illustrate the procedures, they are applied to an n(+)-p solar cell having substrate resistivity of about 0.1 ohm-cm.

Lindholm, F. A.↗
NASA NTRS1984

Crop moisture estimation over the southern Great Plains with dual polarization 1.66 centimeter passive microwave data from Nimbus 7

Moisture content of snow-free, unfrozen soil is inferred using passive microwave brightness temperatures from the scanning multichannel microwave radiometer (SMMR) on Nimbus-7. Investigation is restricted to the two polarizations of the 1.66 cm wavelength sensor. Passive microwave estimates of soil moisture are of two basic categories; those based upon soil emissivity and those based upon the polarization of soil emission. The two methods are compared and contrasted through the investigation of 54 potential functions of polarized brightness temperatures and, in some cases, ground-based temperature measurements. Of these indices, three are selected for the estimated emissivity, the difference between polarized brightness temperatures, and the normalized polarization difference. Each of these indices is about equally effective for monitoring soil moisture. Using an antecedent precipitation index (API) as ground control data, temporal and spatial analyses show that emissivity data consistently give slightly better soil moisture estimates than depolarization data. The difference, however, is not statistically significant. It is concluded that polarization data alone can provide estimates of soil moisture in areas where the emissivity cannot be inferred due to nonavailability of surface temperature data.

Mcfarland, M. J.↗
NASA NTRS1993

An improved gravity model for Mars: Goddard Mars Model 1

Doppler tracking data of three orbiting spacecraft have been reanalyzed to develop a new gravitational field model for the planet Mars, Goddard Mars Model 1 (GMM-1). This model employs nearly all available data, consisting of approximately 1100 days of S band tracking data collected by NASA's Deep Space Network from the Mariner 9 and Viking 1 and Viking 2 spacecraft, in seven different orbits, between 1971 and 1979. GMM-1 is complete to spherical harmonic degree and order 50, which corresponds to a half-wavelength spatial resolution of 200-300 km where the data permit. GMM-1 represents satellite orbits with considerably better accuracy than previous Mars gravity models and shows greater resolution of identifiable geological structures. The notable improvement in GMM-1 over previous models is a consequence of several factors: improved computational capabilities, the use of otpimum weighting and least squares collocation solution techniques which stabilized the behavior of the solution at high degree and order, and the use of longer satellite arcs than employed in previous solutions that were made possible by improved force and measurement models. The inclusion of X band tracking data from the 379-km altitude, nnear-polar orbiting Mars Observer spacecraft should provide a significant improvement over GMM-1, particularly at high latitudes where current data poorly resolve the gravitational signature of the planet.

Smith, D. E.↗
NASA NTRS2000

Evaluation of Mars Entry Reconstructured Trajectories Based on Hypothetical 'Quick-Look' Entry Navigation Data

A first order simulation analysis of the navigation accuracy expected from various Navigation Quick-Look data sets is performed. Here quick-look navigation data are observations obtained by hypothetical telemetried data transmitted on the fly during a Mars probe's atmospheric entry. In this simulation study, navigation data consists of 3-axis accelerometer sensor and attitude information data. Three entry vehicle guidance types are studied: I. a Maneuvering entry vehicle (as with Mars 01 guidance where angle of attack and bank angle are controlled); II. Zero angle-of-attack controlled entry vehicle (as with Mars 98); and III. Ballistic, or spin stabilized entry vehicle (as with Mars Pathfinder);. For each type, sensitivity to progressively under sampled navigation data and inclusion of sensor errors are characterized. Attempts to mitigate the reconstructed trajectory errors, including smoothing, interpolation and changing integrator characteristics are also studied.

Pastor, P. Rick↗
NASA NTRS2016

Characterization of Carbon Dioxide Washout Measurement Techniques in the Mark-III Space Suit

A space suit must provide adequate carbon dioxide (CO2) washout inside the helmet to prevent symptoms of hypercapnia. In the past, an oronasal mask has been used to measure the inspired air of suited subjects to determine a space suit's CO2 washout capability. While sufficient for super-ambient pressure testing of space suits, the oronasal mask fails to meet several human factors and operational criterion needed for future sub-ambient pressure testing (e.g. compatibility with a Valsalva device). This paper describes the evaluation of a nasal cannula as a device for measuring inspired air within a space suit. Eight test subjects were tasked with walking on a treadmill or operating an arm ergometer to achieve target metabolic rates of 1000, 2000, and 3000 British thermal units per hour (BTU/hr), at flow rates of 2, 4, and 6 actual cubic feet per minute (ACFM). Each test configuration was conducted twice, with subjects instructed to breathe either through their nose only, or however they felt comfortable. Test data shows that the nasal cannula provides more statistically consistent data across test subjects than the oronasal mask used in previous tests. The data also shows that inhaling/exhaling through only the nose provides a lower sample variance than a normal breathing style. Nose-only breathing reports better CO2 washout due to several possible reasons, including a decreased respiratory rate, an increased tidal volume, and because nose-only breathing directs all of the exhaled CO2 down and away from the oronasal region. The test subjects in this study provided feedback that the nasal cannula is comfortable and can be used with the Valsalva device.

Meginnis, Ian M.↗
NASA NTRS1992

A P-band SAR interference filter

The synthetic aperture radar (SAR) interference filter is an adaptive filter designed to reduce the effects of interference while minimizing the introduction of undesirable side effects. The author examines the adaptive spectral filter and the improvement in processed SAR imagery using this filter for Jet Propulsion Laboratory Airborne SAR (JPL AIRSAR) data. The quality of these improvements is determined through several data fidelity criteria, such as point-target impulse response, equivalent number of looks, SNR, and polarization signatures. These parameters are used to characterize two data sets, both before and after filtering. The first data set consists of data with the interference present in the original signal, and the second set consists of clean data which has been coherently injected with interference acquired from another scene.

Taylor, Victor B.↗
DOE OSTI2024

Climate-invariant machine learning

Projecting climate change is a generalization problem: We extrapolate the recent past using physical models across past, present, and future climates. Current climate models require representations of processes that occur at scales smaller than model grid size, which have been the main source of model projection uncertainty. Recent machine learning (ML) algorithms hold promise to improve such process representations but tend to extrapolate poorly to climate regimes that they were not trained on. To get the best of the physical and statistical worlds, we propose a framework, termed “climate-invariant” ML, incorporating knowledge of climate processes into ML algorithms, and show that it can maintain high offline accuracy across a wide range of climate conditions and configurations in three distinct atmospheric models. Our results suggest that explicitly incorporating physical knowledge into data-driven models of Earth system processes can improve their consistency, data efficiency, and generalizability across climate regimes.

54 ENVIRONMENTAL SCIENCES↗