Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Input Sensitivity”

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 577 records · Page 32
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Uncertainty Quantification and Sensitivity Analysis in Process-Structure-Property Simulations for Laser Powder Bed Fusion Additive Manufacturing

Process variations and process-induced defects like porosity cause significant uncertainty in the microstructure and mechanical behavior of additively manufactured metals. Establishing process-structure-property (PSP) relationships and quantifying uncertainty using experiments alone is costly, especially for structural applications where mechanical allowables must be established for qualification and certification. This work presents a PSP simulation framework for laser powder bed fusion with a focus on uncertainty quantification through probabilistic calibration and multi-fidelity uncertainty propagation. Motivated by phenomenological input parameters related to grain nucleation and growth that are difficult to characterize, a global sensitivity analysis (GSA) is completed. Through GSA, the most important input parameters are identified based on their influence on the statistical distributions of microstructural metrics that influence mechanical behavior, including grain size, morphology, and crystallographic texture. The results provide insight on what experiments are necessary to quantify and control PSP uncertainties, particularly those associated with the more challenging input parameters.

additive manufacturing↗
NASA NTRS1984

New Uses for Sensitivity Analysis: How Different Movement Tasks Effect Limb Model Parameter Sensitivity

Original results for a newly developed eight-order nonlinear limb antagonistic muscle model of elbow flexion and extension are presented. A wider variety of sensitivity analysis techniques are used and a systematic protocol is established that shows how the different methods can be used efficiently to complement one another for maximum insight into model sensitivity. It is explicitly shown how the sensitivity of output behaviors to model parameters is a function of the controller input sequence, i.e., of the movement task. When the task is changed (for instance, from an input sequence that results in the usual fast movement task to a slower movement that may also involve external loading, etc.) the set of parameters with high sensitivity will in general also change. Such task-specific use of sensitivity analysis techniques identifies the set of parameters most important for a given task, and even suggests task-specific model reduction possibilities.

Winters, J. M.↗
NASA NTRS2013

Simple Sensitivity Analysis for Orion GNC

The performance of Orion flight software, especially its GNC software, is being analyzed by running Monte Carlo simulations of Orion spacecraft flights. The simulated performance is analyzed for conformance with flight requirements, expressed as performance constraints. Flight requirements include guidance (e.g. touchdown distance from target) and control (e.g., control saturation) as well as performance (e.g., heat load constraints). The Monte Carlo simulations disperse hundreds of simulation input variables, for everything from mass properties to date of launch.We describe in this paper a sensitivity analysis tool (Critical Factors Tool or CFT) developed to find the input variables or pairs of variables which by themselves significantly influence satisfaction of requirements or significantly affect key performance metrics (e.g., touchdown distance from target). Knowing these factors can inform robustness analysis, can inform where engineering resources are most needed, and could even affect operations. The contributions of this paper include the introduction of novel sensitivity measures, such as estimating success probability, and a technique for determining whether pairs of factors are interacting dependently or independently. The tool found that input variables such as moments, mass, thrust dispersions, and date of launch were found to be significant factors for success of various requirements. Examples are shown in this paper as well as a summary and physics discussion of EFT-1 driving factors that the tool found.

Monte Carlo Simulation↗
DOE OSTI2022

Understanding radiation effects in friction stir welded MA956 using ion irradiation and a rate theory model

We report an outstanding challenge in the manufacturing and joining of oxide dispersion strengthened steels is retaining the nanofeatures in the alloy throughout the fabrication and welding process. MA956 was friction stir welded with two different sets of welding parameters, resulting in a medium and high heat input. After welding, 5 MeV Fe ++ ion irradiations were performed at doses ranging from 50 to 200 dpa in the temperature range of 400 to 500°C. Post-irradiation characterization was performed with scanning transmission electron microscopy and energy-dispersive x-ray spectroscopy to investigate the Y-Al-O dispersoids, voids, and dislocations. After welding, the dispersoid microstructure coarsened, resulting in fewer and larger dispersoids regardless of heat input. After irradiation, the dispersoid behavior in the welded material was sensitive to temperature, exhibiting growth behavior attributed to Ostwald coarsening at 500 °C but a mixture of nucleation and more muted growth at 400 and 450 °C, attributed to competing mechanisms of radiation-enhanced diffusion and Ostwald coarsening. Void swelling correlated to heat input; being more prevalent in the welded conditions occurring at lower doses and in higher values relative to the base material. The low values of swelling despite microstructure coarsening caused by welding demonstrate the excellent swelling resistance of MA956, even after welding with the highest swelling values of 0.5% noted in the stir zone high heat input condition at 450 °C, 200 dpa. The dislocation behavior was inconsistent: the strongest trend was that network density was higher for welded versus base material, and an increase in loop diameter with temperature was observed. A rate theory model based on the observed microstructure suggests at high temperature interstitial loss to sinks was more likely to be dominant compared to mutual annihilation via point defect recombination, because of an increase of the radiation diffusion coefficient with temperature regardless of initial welded microstructure.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Laying a Foundation for Performance Assessment (PA) Model Development - 19267

Performance Assessment (PA) models are abstractions of real-world conditions. As abstractions, modelers must make a number of decisions regarding model complexity and which features, events, and processes (FEPs) will be represented. To ensure that such decisions are transparent to stakeholders, modelers should document their approach and how this approach influences conceptual model development. This paper describes one such approach. Examples are used to describe an analysis of the potential FEPs. Through the FEPs analysis, all potentially relevant FEPs can be identified. These potentially relevant FEPs are then used to develop a detailed conceptual model known as the Central Scenario, which includes a description of the assessment context, the key features of the site and the closure system, and the expected processes of the system. From this Central Scenario conceptual model, mathematical models can be developed and implemented within modeling software and used to demonstrate compliance to PA performance objectives. FEPs which are beyond the scope of the Central Scenario are identified and used as the basis for developing alternative conceptual models (ACMs) or sensitivity analyses. Once these models and modeling cases are defined, the necessary inputs can be identified and defined. (author)

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Probing Coherent States of Light and Matter in Two-Dimensional Semiconductors. Final Report

The ability to enhance light-matter interactions in engineered optical environments is well-established in micro- and nano-photonics. New classes of materials bring new rich correlations between spin, momentum, and light polarization that can be exploited in these photonic systems for information processing and low-power electronics. Interfacing these novel material features with strong optical interactions suggests the compelling capability to create hybrid light-matter systems harnessing low-energy and protected material properties such as spin. The long-term objective of this program was to investigate the interplay between pseudospin in 2D materials and light, revealing new coherent phenomena. The materials studied are the monolayer transition metal dichalcogenides (TMDs), which support polarization-sensitive optical transitions when isolated to a single sub-nanometer crystal layer in thickness. In particular, monolayer TMD semiconductors such as MoS 2 exhibit degenerate valleys in momentum space with distinct spin character. These valleys can be separately addressed by circularly polarized light, providing a tool for manipulation. Integrating this valley pseudospin of excitons in monolayer TMDs with engineered strong optical excitations to create novel coherent phenomena was a key long-term goal of this research. The scope of this research primarily focused on studying light-matter interactions of valley-sensitive excitations in monolayer TMDs. The scope also included the integration of monolayer materials that support valley-polarized emission into photonic devices. This approach extends prior work on exciton-polaritons in 2D materials toward a platform with far more control and improved light input and output characteristics to aid scalability of these polarization-sensitive phenomena to many-body, long-range optical networks. During the course of the five-year Early Career program, the approaches led to successful progress on each proposed aim. Initial successful work seeded additional research that developed from the main themes, while this program focus evolved toward coherent manipulation of valley polaritons and integration with photonics. Both of these advanced topics set the stage for new capabilities in future innovative research. The results of this project provide advances in materials and methods relevant for new visions of opto-electronics and photonics that can improve information processing in an era of increased focus on quantum and coherent phenomena. The research has also contributed to the training of two postdoctoral scholars and five graduate students.

2D Materials↗
NASA NTRS1975

A study of moving base simulation motion cues utilizing washout technique

The present study was conducted to derive data on non-visual motion thresholds utilizing washout technique, and to develop specific threshold values for use as washout parameters. It describes the results of acceleration detection studies carried out using the NASA MSFC General Purpose Simulator which provides six degree-of-freedom cab motion. A series of experiments was performed to test the hypothesis that constant velocity visual cues might suppress non-visual deceleration sensitivity. The psychophysical method employed was the forced-choice procedure which theoretically yields a pure sensitivity measure. The MSFC General Purpose Simulator was programmed to provide velocity ramps in three axes - pitch, fore-aft translation, and vertical translation. Data were collected with and without a constant velocity visual input. Comparison of the current results with classical data suggested significantly greater sensitivity of the human observer to fore-aft accelerations than has generally been reported, sensitivity tending to depend on motion direction and sign of velocity change.

Kirkpatrick, M., III↗
NASA NTRS2000

Optical Sensors Based on Single on Arm Thin Film Waveguide Interferometer

Single-arm dual-mode optical waveguide interferometer utilizes interference between two modes of different order. Sensing effect results from the change in propagation conditions of the modes caused by the environment. The waveguide is made as an open asymmetric structure containing a dye-doped polymer film onto a quartz substrate. It is more sensitive to the change of environment than its conventional polarimetric analog using orthogonal modes (TE and TM) of the same order. The sensor still preserves the option of operating in polarimetric regime using a variety of mode combinations such as TE(sub 0)/TM(sub 0) (conventional) TE(sub 0)/TM(sub 1), TE(sub 1)/TM(sub 0), or TE(sub 1)/TM(sub 1) but can also work in nonpolarimetric regime using combinations TE(sub 0)/TE(sub 1) or TM(sub 0)/TM(sub 1). Utilization of different mode combinations simultaneously makes the device more versatile. Application of the sensor to gas sensing is based on doping polymer film with an organic indicator dye targeting a particular gaseous reagent. Change of the optical absorption spectrum of the dye caused by the gaseous pollutant results in change of the reactive index of the dye-doped polymer film that can be detected by the sensor. As indicator dyes we utilize Bromocresol Purple doped into polymer poly(methyl) methacrylate that is sensitive to small concentrations of ammonia. The indicator dye demonstrated an irreversible increase in optical absorption near the peak at 350 nm being exposed to 5% ammonia in pure nitrogen at 600 Torr. The dye also showed reversible growth of the absorption peak near 600 nm after exposure to a vapor of standard medical ammonia spirit (65% alcohol). We have built a breadboard prototype of the sensor with He-Ne laser as a light source and with a single mode fiber input and a multimode fiber output. The prototype showed a sensitivity to temperature change of the order of 2 C per 2pi phase shift. The sensitivity of the sensor to the presence of dTy ammonia is not less than 300 ppm per 2pi phase shift. The proposed sensor can be used as a robust stand-alone instrument for continuous environment pollution monitoring.

Sarkisov, S. S.↗
DOE OSTI2023

β -delayed neutron emissions from N > 50 gallium isotopes

β-delayed γ-neutron spectroscopy has been performed on the decay of A = 84 to 87 gallium isotopes at the RI-beam Factory at the RIKEN Nishina Center using a high-efficiency array of 3 He neutron counters (BRIKEN). β-2n-γ events were measured in the decays of all of the four isotopes for the first time, which is direct evidence for populating the excited states of two-neutron daughter nuclei. Detailed decay schemes with the γ branching ratios were obtained for these isotopes, and the neutron emission probabilities (P xn ) were updated from the previous study. Hauser-Feshbach statistical model calculations were performed to understand the experimental branching ratios. We found that the P 1n and P 2n values are sensitive to the nuclear level densities of 1n daughter nuclei and showed that the statistical model reproduced the P 2n /P 1n ratio better when experimental levels plus shell-model level densities fit by the Gilbert-Cameron formula were used as the level-density input. We also showed the neutron and γ branching ratios are sensitive to the ground-state spin of the parent nucleus. Our statistical model analysis suggested J ≤ 3 for the unknown ground-state spin of the odd-odd nucleus 86 Ga, from the I γ (4 + → 2 + )/I γ (2 + → 0 + ) ratio of 84 Ga and the P 2n /P 1n ratio. In conclusion, these results show the necessity of detailed understanding of the decay scheme, including data from neutron spectroscopy, in addition to γ measurements of the multineutron emitters.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2018

On the Usage of Ultrasound Computational Models for Decision Making Under Ambiguity

Computer modeling and simulation is becoming pervasive within the non-destructive evaluation (NDE) industry as a convenient tool for designing and assessing inspection techniques. This raises a pressing need for developing quantitative techniques for demonstrating the validity and applicability of the computational models. Computational models provide deterministic results based on deterministic and well-defined input, or stochastic results based on inputs defined by probability distributions. However, computational models cannot account for the effects of personnel, procedures, and equipment, resulting in ambiguity about the efficacy of inspections based on guidance from computational models only. In addition ambiguity arises when model inputs, such as the representation of realistic cracks, cannot be defined deterministically, probabilistically, or by intervals. In this work, Pacific Northwest National Laboratory demonstrates the ability of computational models to represent field measurements under known variabilities, and quantify the differences using maximum amplitude and power spectrum density metrics. Sensitivity studies are also conducted to quantify the effects of different input parameters on the simulation results.

Dib, Gerges↗
DOE OSTI2023

Sensitivity Analysis of a Polyphase Wireless Power Transfer System under Off-Nominal Conditions

Misaligned and/or variable coil airgaps cause coupling coefficient variation in wireless power transfer (WPT) systems, resulting in a decrease in the system’s transmitted power and efficiency. This paper presents sensitivity analyses of a three-phase, Y-Y connected, series-tuned WPT system in the frequency domain in terms of several different electric vehicle wireless charging off-nominal conditions (misalignments in Δ x and Δ y directions, change of airgap Δ z , and the roll Δψ, pitch Δθ, and yaw ΔΦ angles) as specified in Society of Automotive Engineers (SAE) J2954 Standard. Coil inductance matrices were obtained by measuring the self- and mutual inductances of the primary and secondary coil phase windings at variable airgap classes (from 5 cm to 30 cm) for five different charging positions as identified in SAE J2954. These 6 × 6 inductance matrices were used in sensitivity and Plexim/PLECS simulation analyses. The sensitivity of the WPT system was analyzed analytically using the input impedance and phase angle, voltage gain, current gain, quality factor, and coupling coefficient parameters of the series-tuned WPT system. The results were confirmed experimentally on a 50 kW WPT system.

42 ENGINEERING↗
DOE OSTI2022

Challenges and Lessons Learned in Applying Sensitivity Analysis to Building Stock Energy Models

Uncertainty Analysis (UA) and Sensitivity Analysis (SA) offer essential tools to determine the limits of inference of a model and explore the factors which have the most effect on the model outputs. However, despite a well established body of work applying UA and SA to models of individual buildings, a review of the literature relating to energy models for larger groups of buildings undertaken by Fennell et al. (2019) highlighted very limited application at larger scales. This contribution describes the efforts undertaken by a group of research teams in the context of IEA-EBC Annex 70 working with a diverse set of Building Stock Models (BSMs) to apply global sensitivity analysis methods and compare their results. Since BSMs are a class of model defined by their output and coverage rather than their structure and inputs, they represent a diverse set of modelling approaches. Key challenges for the application of SA are identified and explored, including the influence of model form, input data types and model outputs. This study combines results from 7 different modelling teams, each using different models across a range of urban areas to explore these challenges and begin the process of developing standardised workflows for SA of BSMs.

ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATION↗
DOE OSTI2019

Development of a Stochastic Infiltration Model for PPA Implementation - 19293

Performance Assessment models often require inputs representing specific hydrologic processes and properties such as infiltration, water content, groundwater flow rate, streamflow, and sediment flux. This paper details the development of a process-level infiltration model and its use in a Probabilistic Performance Assessment (PPA) model for the West Valley Site in western New York. Additionally, this paper demonstrates a method for obtaining distributions of the long-term average hydrologic processes and properties for a PA model. For this PPA, an infiltration model was developed using HELP (Hydrologic Evaluation of Landfill Performance) to evaluate the rates of infiltration, percolation, and drainage through several site soil and proposed decommissioning cover systems, as well as the volumetric water contents of each of the unsaturated layers in a modeled soil column. A particular challenge of probabilistic modeling over long periods of time is scaling parameters appropriately so that outcomes are not grossly over- or underestimated. For these models, distributions of climatic parameters such as average precipitation rate, temperature, wind speed, humidity, and solar radiation were scaled to 1000-year averages and were used to force steady-state models that predict 1000-year average infiltration rates, drainage rates, and water contents. For the HELP model, distributions were developed for the physical and hydrologic properties of each material in a given soil column or cover system, and these properties were allowed to vary stochastically along with climatic inputs. The HELP model was run 5000 times for each of seven different soil and cover systems representing various existing and proposed conditions (i.e. native soil, backfill, existing covers, and a proposed close-in-place cover). The results of the HELP model provide ranges of 1000-year average infiltration rates, drainage rates, and water contents through and within the various layers of each of the different soil and cover conditions. The model outcomes also demonstrate the effectiveness of a proposed close-in-place cover, which essentially eliminates infiltration during its operational lifetime. Following model simulation, a sensitivity analysis was performed to establish the relative importance of each stochastic input with respect to each output. The outcomes of the model simulations and the sensitivity analysis were implemented into a system-level GoldSim model via a novel sensitivity-informed nearest neighbor approach. These infiltration rates, drainage rates, and water contents are used to parameterize mixing cells within the GoldSim Contaminant Transport module, which is the primary system-level model used in the PPA. The infiltration rates are also used as boundary conditions for a stochastic groundwater flow model that is also utilized in the PPA, but not discussed in this paper. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2016

Hotspot electron temperature from x-ray continuum measurements on the NIF

We report on measurements of the electron temperature in the hotspot of inertially confined, layered, spherical implosions on the National Ignition Facility using a differential filtering diagnostic. Measurements of the DT and DD ion temperatures using neutron time-of-flight detectors are complicated by the contribution of hot spot motion to the peak width, which produce an apparent temperature higher than the thermal temperature. The electron temperature is not sensitive to this non-thermal velocity and is thus a valuable input to interpreting the stagnated hot spot conditions. Here we show that the current differential filtering diagnostic provides insufficient temperature resolution for the hot spot temperatures of interest. We then propose a new differential filter configuration utilizing larger pinhole size to increase spectral fluence, as well as thicker filtration. This new configuration will improve measurement uncertainty by more than a factor of three, allowing for a more accurate hotspot temperature.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2018

Life cycle greenhouse gas emissions of hydrogen fuel production from chlor-alkali processes in the United States

By-product hydrogen from chlor-alkali processes can help meet the increasing demand for hydrogen fuel in early fuel cell electric vehicle markets (e.g., California) in the U.S. Hydrogen produced from chlor-alkali plants is typically combusted for process heat on site, vented to the atmosphere (i.e., wasted), or sold to the external merchant hydrogen market. Whether it is combusted, vented, or sold as a commodity, relevant information is lacking as to the life-cycle environmental benefits or trade-offs of using by-product hydrogen from chlor-alkali plants. A life-cycle analysis framework was employed to evaluate well-to-gate greenhouse gas (GHG) emissions associated with by-product hydrogen from chlor-alkali processes in comparison with hydrogen from the conventional centralized natural gas steam methane reforming (central SMR) pathway. U.S.-specific, plant-by-plant, and up-to-date chlor-alkali production characteristics were incorporated into the analysis. In addition to the venting and combustion scenarios, to deal with the multi-functionality of the chlor-alkali processes that simultaneously produce chlorine, sodium hydroxide, and hydrogen, two different co-product allocation strategies were adopted mass allocation and market value allocation. It was estimated that by-product hydrogen production from chlor-alkali processes creates 1.3-9.8 kg CO 2 e/kg H 2 of life-cycle GHG emissions on average, which is 20-90% less than the conventional central SMR pathway. The results vary with co-product treatment scenarios, regional electric grid characteristics, on-site power generation, product prices, and hydrogen yield. Despite the variations in the results, it was concluded that the life-cycle GHG emission reduction benefits of using by-product hydrogen from chlor-alkali processes are robust. In conclusion, with a diverse set of scenario analyses, the study developed a comprehensive and detailed life-cycle GHG emissions inventory of the chlor-alkali by-product hydrogen pathway and quantified sensitivity indices in the context of different assumptions and input parameter values.

08 HYDROGEN↗
DOE OSTI2021

25-Pin metallic fuel performance benchmark case based on the EBR-II X430 experiments series

A metallic fuel benchmark case was developed for the fuel performance code BISON based on 25 uranium-zirconium and uranium-plutonium-zirconium pins of the Experimental Breeder Reactor II X430 experiment series. Results of the benchmarks were compared with measurements and calculations made at the time of the experiment as well as subsequent measurements reported in 2019. The comparisons were used to quantify the accuracy of the BISON predictions and to identify patterns in the BISON differences.BISON predicted burnup, plenum pressure, and fission gas release fractions accurately. BISON temperature predictions were somewhat cooler than the temperatures determined at the time of the experiment but appeared to be reasonably accurate considering uncertainties in the legacy temperature calculations, uncertainties in the legacy linear heat rate calculations, and the high sensitivities of the BISON-predicted temperatures to BISON inputs. Fuel axial elongation predictions had errors correlated to fuel composition; BISON tended to underpredict the elongation of binary fuels and overpredict the elongation of ternary fuels. BISON cladding radial dilation predictions were also significantly lower than legacy PIE measurements.Recommendations were made to improve the BISON fuel gaseous swelling model to account for fuel composition, to add additional capabilities to the coolant channel temperature model to ease benchmark development, and to continue developing benchmark cases based on a wide range of experiments in several reactors. Once a wide array of benchmarks is developed, an attempt can be made to enhance or calibrate BISON models to improve the cladding dilation predictions.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2020

Global Techno-Economic Performance of Bifacial and Tracking Photovoltaic Systems

Although most of the current photovoltaic (PV) system installations use monofacial modules with fixed-tilt mounting structures, bifacial modules, trackers, and the combination of them are also getting more attention as potential candidates to further reduce the PV levelized cost of electricity (LCOE). This work is then the first to present a worldwide analysis on the yield potential and cost-effectiveness of PV farms composed of monofacial fixed-tilt and single/dual (1T/2T) tracker installations, as well as their bifacial counterparts. Our approach starts by estimating the irradiance reaching the front and rear surface of the modules for the different system designs (validated based on data from real PV systems and results from the literature) to estimate their energy production (row-row shading is neglected). Subsequently, the overall system cost during their 25-year lifetime is factored in, and LCOE is obtained. The results reveal that bifacial-1T installations increase energy yield by 35% and reach the lowest LCOE for the majority of the world (93.1% of the land area). Although dual axis trackers achieve the highest energy generation — especially for bifacial modules — their cost is still too high, and are therefore not as cost-effective. Sensitivity analyses based on the Monte Carlo and region sensitivity approach are performed to analyze the impact of input assumptions on the calculated LCOE. These reveal that our conclusions are robust in general but exact configuration to choose depends on specific site conditions. This investigation is not only of interest to the scientific community, but also can be used as a guide for PV installation companies and investors to determine the most suitable technology for a particular location.

14 SOLAR ENERGY↗
DOE OSTI2024

Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains

Abstract In this study we construct and compare three different deep learning (DL) models for estimating daily snow water equivalent (SWE) from high‐resolution gridded meteorological fields over the Rocky Mountain region. To train the DL models, Snow Telemetry (SNOTEL) station‐based SWE observations are used as the prediction target. All DL models produce higher median Nash‐Sutcliffe Efficiency (NSE) values than a conceptual SWE model and interpolated gridded data sets, although mean squared errors also tend to be higher. Sensitivity of the SWE prediction to the model's input variables is analyzed using an explainable artificial intelligence (XAI) method, yielding insight into the physical relationships learned by the models. This method reveals the dominant role precipitation and temperature play in snowpack dynamics. In applying our models to estimate SWE throughout the Rocky Mountains, an extrapolation problem arises since the statistical properties of SWE (e.g., annual maximum) and geographical properties of individual grid points (e.g., elevation) differ from the training data. This problem is solved by normalizing the SWE with its historical maximum value to alleviate extrapolation for all tested DL models. Our work shows that the DL models are promising tools for estimating SWE, and sufficiently capture relevant physical relationships to make them useful for spatial and temporal extrapolation of SWE values.

54 ENVIRONMENTAL SCIENCES↗