Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “error estimation and analysis”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,585 records · Page 88
All sourcesNASA NTRSDOE OSTI
NASA NTRS1984

Analysis of the quality of image data acquired by the LANDSAT-4 thematic mapper and multispectral scanners

A seven step procedure developed for evaluating the geometric properties of MSS and TM film produces is being implemented. Some 476 control points were selected of which 238 are being tested and edited for digitization and scaling errors. Tables show statistics established for assessing the spectral characteristics and variability, as well as the spatial resolution and radiometric sensitivity of TM data for a forest environment in an effort to determine the extent to which major forest cover type can be detected and identified on TM digital and image products. Results thus far show that the high quality obtained are more than sufficient for meeting most of the inventory objectives of the renewable resource specialist. The TM data should be extremely valuable for: (1) estimating forest cover types; (2) updating land use survey maps; and (3) determining the size and shape and location of individual forest clearings and water resources.

Colwell, R. N.↗
DOE OSTI2020

Estimating Atmospheric Parameters of DA White Dwarf Stars with Deep Learning

The determination of atmospheric parameters of white dwarf stars (WDs) is crucial for researches on them. Traditional methodology is to fit the model spectra to observed absorption lines and report the parameters with the lowest χ 2 error, which strongly relies on theoretical models that are not always publicly accessible. In this work, we construct a deep learning network to model-independently estimate T eff and log g of DA stars (DAs), corresponding to WDs with hydrogen-dominated atmospheres. The network is directly trained and tested on the normalized flux pixels of full optical wavelength range of DAs spectroscopically confirmed in the Sloan Digital Sky Survey. Experiments in test yield that the rms error for T eff and log g approaches 900 K and 0.1 dex, respectively. This technique is applicable for those DAs with T eff from 5000 to 40,000 K and log g from 7.0 to 9.0 dex. Furthermore, the applicability of this method is verified for the spectra with degraded resolution of ~200. So it is also practical for the analysis of DAs that will be detected by the Chinese Space Station Telescope.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2011

Alaska: Glaciers of Kenai Fjords National Park and Katmai and Lake Clark National Parks and Preserve

There are hundreds of glaciers in Kenai Fjords National Park (KEFJ) and Katmai National Park and Preserve (KATM) covering over 2276 sq km of park land (circa 2000). There are two primary glacierized areas in KEFJ -- the Harding Icefield and the Grewingk-Yalik Glacier Complex, and three primary glacierized areas in KATM - the Mt. Douglas area, the Kukak Volcano to Mt. Katmai area and the Mt. Martin area. Most glaciers in these parks terminate on land, though a few terminate in lakes. Only KEFJ has tidewater glaciers, which terminate in the ocean. Glacier mapping and analysis of the change in glacier extent has been accomplished on a decadal scale using satellite imagery, primarily Landsat data from the 1970s, 1980s, and from 2000. Landsat Multispectral Scanner (MSS), Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) imagery was used to map glacier extent on a park-wide basis. Classification of glacier ice using image processing software, along with extensive manual editing, was employed to create Geographic Information System (GIS) outlines of the glacier extent for each park. Many glaciers that originate in KEFJ but terminate outside the park boundaries were also mapped. Results of the analysis show that there has been a reduction in the amount of glacier ice cover in the two parks over the study period. Our measurements show a reduction of approximately 21 sq km, or -1.5% (from 1986 to 2000), and 76 sq km, or -7.7% (from 1986/87 to 2000), in KEFJ and KATM, respectively. This work represents the first comprehensive study of glaciers of KATM. Issues that complicate the mapping of glacier extent include: debris-cover (moraine and volcanic ash), shadows, clouds, fresh snow, lingering snow from the previous season, and differences in spatial resolution between the MSS and TM or ETM+ sensors. Similar glacier mapping efforts in western Canada estimate mapping errors of 3-4%. Measurements were also collected from a suite of glaciers in KEFJ and KATM detailing terminus positions and rates of recession using datasets including the 15-minute USGS quadrangle maps (1950/1951), Landsat imagery (1986/1987, 2000, 2006) and 2005 Ikonos imagery (KEFJ only).

Giffen, bruce A.↗
DOE OSTI2025

Analysis of RF Surface Loss in a Planar 2D Qubit

The Josephson junction and shunt capacitor form a transmon qubit, which is the cornerstone of modern quantum computing platforms. For reliable quantum computing, it is important how long a qubit can remain in a superposition of quantum states, which is determined by the coherence time (T1). The coherence time of a qubit effectively sets the "lifetime" of usable quantum information, determining how long quantum computations can be performed before errors occur and information is lost. There are several sources of decoherence in transmon qubits, but the predominant one is generally considered to be dielectric losses in the natural oxide layer formed on the surface of the superconductor. In this paper, we present a numerical study of microwave surface losses in planar superconducting antennas of different transmon qubit designs. An asymptotic method for estimating the energy participation ratio in ultrathin films of nanometer scales is proposed, and estimates are given for the limits of achievable minimum RF losses depending on the electrical properties of the surface oxide and the interface of the qubit with the substrate material.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2025

Machine learning and process-based modeling of spatiotemporal changes in active layer thickness across Alaska

Permafrost degradation poses a growing threat to infrastructure stability and ecosystem resilience in the rapidly warming Arctic. We investigated the spatiotemporal dynamics of active layer thickness (ALT) across Alaska by integrating field observations, environmental datasets, a physically based Stefan model, and machine learning (ML) techniques. Using weather projections from the Coupled Model Intercomparison Project Phase 6 under two Shared Socioeconomic Pathways (SSP 2-4.5 and SSP 5-8.5), we assessed ALT sensitivity to projected future weather conditions. The random forest (RF) model outperformed the Stefan approach in predicting ALT on the training dataset (R² = 0.84 vs. 0.53) but demonstrated lower generalizability on the test dataset (R² = 0.24 vs. 0.54). The root mean square error (RMSE) for the RF model for training and testing ranged from 14 to 22 cm, compared to 17 and 18 cm for the Stefan model. Variable importance analysis revealed that mean annual temperature and slope angle were the strongest predictors of ALT, accounting for 19% and 18% of the variance, respectively, followed by sediment transport index (14%) and stream power index (11%). Comparative analysis of baseline ALT predictions showed the Stefan model tended to project a thicker active layer (mean ± SD: 65 ± 16 cm), compared to the RF model (mean ± SD: 59 ± 8.8) cm). Both models indicated a latitudinal gradient in ALT, with shallower depths at higher latitudes. Projected ALT increases by 2100 were estimated at 3.3 ± 2.2 cm under SSP 2-4.5 and 5.9 ± 4.0 cm under SSP 5-8.5 for the ML model, whereas the Stefan model projected substantially larger increases of 13 ± 2.6 cm (SSP 2-4.5) and 28 ± 4.4 cm (SSP5-8.5). Spatial analysis showed the greatest ALT increases in northern Alaska, with relatively smaller changes in southern regions. These findings highlight the complex, multifactorial nature of ALT dynamics and the value of hybrid modeling approaches. As rising temperatures accelerate permafrost thaw, changes in ALT can disrupt ecosystems, damage infrastructures, and enhance the release of stored soil carbon, highlighting the urgent need for improved predictive capabilities to inform adaptation strategies in the Arctic.

Climate sciences↗
NASA NTRS1993

Simulation of wave propagation in three-dimensional random media

Quantitative error analysis for simulation of wave propagation in three dimensional random media assuming narrow angular scattering are presented for the plane wave and spherical wave geometry. This includes the errors resulting from finite grid size, finite simulation dimensions, and the separation of the two-dimensional screens along the propagation direction. Simple error scalings are determined for power-law spectra of the random refractive index of the media. The effects of a finite inner scale are also considered. The spatial spectra of the intensity errors are calculated and compared to the spatial spectra of intensity. The numerical requirements for a simulation of given accuracy are determined for realizations of the field. The numerical requirements for accurate estimation of higher moments of the field are less stringent.

Coles, William A.↗
NASA NTRS2010

Simulation Studies for a Space-Based CO2 Lidar Mission

We report results of initial space mission simulation studies for a laser-based, atmospheric CO2 sounder, which are based on real-time carbon cycle process modelling and data analysis. The mission concept corresponds to the Active Sensing of CO2 Emissions over Nights, Days and Seasons (ASCENDS) recommended by the US National Academy of Sciences' Decadal Survey. As a pre-requisite for meaningful quantitative evaluation, we employ a CO2 model that has representative spatial and temporal gradients across a wide range of scales. In addition, a relatively complete description of the atmospheric and surface state is obtained from meteorological data assimilation and satellite measurements. We use radiative transfer calculations, an instrument model with representative errors and a simple retrieval approach to quantify errors in 'measured' CO2 distributions, which are a function of mission and instrument design specifications along with the atmospheric/surface state. Uncertainty estimates based on the current instrument design point indicate that a CO2 laser sounder can provide data consistent with ASCENDS requirements and will significantly enhance our ability to address carbon cycle science questions. Test of a dawn/dusk orbit deployment, however, shows that diurnal differences in CO2 column abundance, indicative of plant photosynthesis and respiration fluxes, will be difficult to detect

Kawa, S. R.↗
NASA NTRS2006

Excitations for Rapidly Estimating Flight-Control Parameters

A flight test on an F-15 airplane was performed to evaluate the utility of prescribed simultaneous independent surface excitations (PreSISE) for real-time estimation of flight-control parameters, including stability and control derivatives. The ability to extract these derivatives in nearly real time is needed to support flight demonstration of intelligent flight-control system (IFCS) concepts under development at NASA, in academia, and in industry. Traditionally, flight maneuvers have been designed and executed to obtain estimates of stability and control derivatives by use of a post-flight analysis technique. For an IFCS, it is required to be able to modify control laws in real time for an aircraft that has been damaged in flight (because of combat, weather, or a system failure). The flight test included PreSISE maneuvers, during which all desired control surfaces are excited simultaneously, but at different frequencies, resulting in aircraft motions about all coordinate axes. The objectives of the test were to obtain data for post-flight analysis and to perform the analysis to determine: 1) The accuracy of derivatives estimated by use of PreSISE, 2) The required durations of PreSISE inputs, and 3) The minimum required magnitudes of PreSISE inputs. The PreSISE inputs in the flight test consisted of stacked sine-wave excitations at various frequencies, including symmetric and differential excitations of canard and stabilator control surfaces and excitations of aileron and rudder control surfaces of a highly modified F-15 airplane. Small, medium, and large excitations were tested in 15-second maneuvers at subsonic, transonic, and supersonic speeds. Typical excitations are shown in Figure 1. Flight-test data were analyzed by use of pEst, which is an industry-standard output-error technique developed by Dryden Flight Research Center. Data were also analyzed by use of Fourier-transform regression (FTR), which was developed for onboard, real-time estimation of the derivatives.

Moes, Tim↗
DOE OSTI2020

Effective Opacity of the Intergalactic Medium from Galaxy Spectra Analysis

We measure the effective opacity (τ{sub eff}) of the intergalactic medium from the composite spectra of 281 Lyman-break galaxies in the redshift range 2 ≲ z ≲ 3. Our spectra are taken from the COSMOS Lyα Mapping And Tomographic Observations survey derived from the Low Resolution Imaging Spectrometer on the W.M. Keck I telescope. We generate composite spectra in two redshift intervals and fit them with spectral energy distribution (SED) models composed of simple stellar populations. Extrapolating these SED models into the Lyα forest, we measure the effective Lyα opacity (τ{sub eff}) in the 2.02 ≤ z ≤ 2.44 range. At z = 2.22, we estimate τ{sub eff} =0.159±0.001 from a power-law fit to the data. These measurements are consistent with estimates from quasar analyses at z < 2.5 indicating that the systematic errors associated with normalizing quasar continua are not substantial. We provide a Gaussian processes model of our results and previous τ{sub eff} measurements that describes the steep redshift evolution in τ{sub eff} from z = 1.5–4.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Velocity reconstruction in the era of DESI and Rubin/LSST. I. Exploring spectroscopic, photometric, and hybrid samples

Peculiar velocities of galaxies and halos can be reconstructed from their spatial distribution alone. This technique is analogous to the baryon acoustic oscillations reconstruction, using the continuity equation to connect density and velocity fields. The resulting reconstructed velocities can be used to measure imprints of galaxy velocities on the cosmic microwave background like the kinematic Sunyaev-Zel’dovich effect or the moving lens effect. As the precision of these measurements increases, characterizing the performance of the velocity reconstruction becomes crucial to allow unbiased and statistically optimal inference. In this paper, we quantify the relevant performance metrics: the variance of the reconstructed velocities and their correlation coefficient with the true velocities. We show that the relevant velocities to reconstruct for kSZ and moving lens are actually the halo—rather than galaxy—velocities. We quantify the impact of redshift-space distortions, photometric redshift errors, satellite galaxy fraction, incorrect cosmological parameter assumptions and smoothing scale on the reconstruction performance. Here, we also investigate hybrid reconstruction methods, where velocities inferred from spectroscopic samples are evaluated at the positions of denser photometric samples. We find that using exclusively the photometric sample is better than performing a hybrid analysis. The 2 Gpc/ℎ length simulations from abacussummit with realistic galaxy samples for DESI and Rubin LSST allow us to perform this analysis in a controlled setting. In the companion paper [B. Hadzhiyska, S. Ferraro, B. Ried Guachalla, and E. Schaan, companion paper, Phys. Rev. D 109, 103534 (2024).], we further include the effects of evolution along the light cone and give realistic performance estimates for DESI luminous red galaxies, emission line galaxies, and Rubin LSST-like samples.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2022

The TRAPPIST-1 Habitable Atmosphere Intercomparison (THAI). Part III: Simulated Observables - The Return of the Spectrum

The TRAPPIST-1 Habitable Atmosphere Intercomparison (THAI) is a community project that aims to quantify how differences in general circulation models (GCMs) could impact the climate prediction for TRAPPIST-le and, subsequently its atmospheric characterization in transit. Four GCMs have participated in THAI so far: ExoCAM, LMD-Generic, ROCKE-3D and the UM. This paper, focused on the simulated observations, is the third part of a trilogy, following the analysis of two land planet scenarios (part I) and two aquaplanet scenarios (part II). Here, we show a robust agreement between the simulated spectra and the number of transits estimated to detect the land planet atmospheres. For the cloudy aquaplanet ones, a 5-σ detection of CO2 could be achieved in about 10 transits if the atmosphere contains at least 1 bar of CO2. That number can vary by 41- 56% depending on the GCM used to predict the terminator profiles, principally due to differences in the cloud deck altitude, with ExoCAM and LMD-G producing higher clouds than ROCKE-3D and UM. Therefore, for the first time, this work provides "GCM uncertainty error bars" of ~ 50% that need to be considered in future analyses of transmission spectra. We also analyzed the inter-transit variability induced by weather patterns and changes of terminator cloudiness between transits. Its magnitude differs significantly between the GCMs but its impact on the transmission spectra is within the measurement uncertainties. THAI has demonstrated the importance of model intercomparison for exoplanets and also paved the way for a larger project to develop an intercomparison meta-framework, namely the Climates Using Interactive Suites of Intercomparisons Nested for Exoplanet Studies (CUISINES).

Exoplanet atmospheres↗
DOE OSTI2026

Biases in preconstruction estimates of wind plant annual energy production

Estimating the energy yield of a wind plant during the preconstruction phase is a historically difficult task, even with industry improvements in these estimations. We build on prior research comparing the realized energy production of wind plants and their estimated annual energy production P50 values (median energy production), using owner-provided energy production and losses. We produced similar results to prior studies but with a slightly increasing bias of overestimating median energy production (a bias between realized and estimated energy production of −7.4 % to −6.6 %, depending on the scenario, as opposed to −6.7 % to −5.5 % from earlier studies). In addition to assessing annual energy production P50 bias, we compared both the 1-year and the long-term annual energy production P90 and uncertainty energy yield assessment estimates to the observed long-term-corrected energy production. We found that neither the energy yield assessment uncertainty nor the P90 is conservative enough compared to the observed distribution of prediction errors, suggesting significant room for improvement in the energy yield assessment process.

17 WIND ENERGY↗
DOE OSTI2023

Refractivity Observations from Radar Phase Measurements: The 22 May 2002 Dryline Case during IHOP Project

The dryline, often associated with the development of severe storms in the Southern Great Plains of the United States of America, is a boundary layer phenomenon that occurs when a warm and moist air mass from the Gulf of Mexico meets a hot and dry air mass from the southwest desert area. An accurate knowledge of the water vapor spatio-temporal variability in the lower part of the atmosphere is crucial for a better understanding of the evolution of the dryline. The tropospheric refractivity, directly related to water vapor content, is a proxy for the water vapor content of the troposphere. It has already been demonstrated that the refractivity and the refractivity vertical gradient can be jointly estimated from radar phase measurements. In fact, it has been shown that using kriging interpolation techniques, accurate refractivity maps within the coverage area of the radar can be obtained with high temporal resolution. In this paper, a detailed analysis of the time series of radar-based refractivity maps obtained during a dryline that occurred on the afternoon of 22 May 2002 during the International H 2 O Project is presented. Comparisons between the time series of radar refractivity maps, obtained with the NCAR S-Pol radar, and the refractivity measurements derived from automatic ground-based weather stations and the AERI instrument, placed at different locations within the coverage area of the NCAR S-Pol radar, demonstrate the accuracy of radar refractivity estimates even for highly variable conditions, both in time and space, in the troposphere. Correlation coefficients higher than 0.95 are obtained in all weather station locations. Regarding the RMSE, errors less than 6 N-units are obtained for all cases, being even as low as 2.92 N-units at some locations.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1993

A model for the beam-filling effect associated with the microwave retrieval of rain

Estimating rain rate from environmental microwave emissions is hampered by several difficulties. One of these difficulties is known as the beam-filling effect. Beam filling is the systematic error introduced when the microwave radiometer's field of view is not filled with uniform rain. Beam filling can have dramatic effects on rain-rate estimation, causing rain rates to be underestimated by as much as a factor of 2. The present study derives an approximate expression for beam filling that provides, in principle, a way to estimate this effect. In addition, this study deals only with single-channel microwave rain estimation over the ocean. The final results reveal that beam filling is essentially determined by the freezing level, the mean fraction of the footprint raining, and the footprint-averaged rain rate. Also, the numerical results appear to agree with other empirical studies. Furthermore, the analysis brings to light an interesting connection with rain threshold techniques for estimating area-averaged rain rates.

Graves, Charles E.↗
NASA NTRS2017

Validation of SMAP Surface Soil Moisture Products with Core Validation Sites

The NASA Soil Moisture Active Passive (SMAP) mission has utilized a set of core validation sites as the primary methodology in assessing the soil moisture retrieval algorithm performance. Those sites provide well calibrated in situ soil moisture measurements within SMAP product grid pixels for diverse conditions and locations.The estimation of the average soil moisture within the SMAP product grid pixels based on in situ measurements is more reliable when location specific calibration of the sensors has been performed and there is adequate replication over the spatial domain, with an up-scaling function based on analysis using independent estimates of the soil moisture distribution. SMAP fulfilled these requirements through a collaborative CalVal Partner program.This paper presents the results from 34 candidate core validation sites for the first eleven months of the SMAP mission. As a result of the screening of the sites prior to the availability of SMAP data, out of the 34 candidate sites 18 sites fulfilled all the requirements at one of the resolution scales (at least). The rest of the sites are used as secondary information in algorithm evaluation. The results indicate that the SMAP radiometer-based soil moisture data product meets its expected performance of 0.04 cu m/cu m volumetric soil moisture (unbiased root mean square error); the combined radar-radiometer product is close to its expected performance of 0.04 cu m/cu m, and the radar-based product meets its target accuracy of 0.06 cu m/cu m (the lengths of the combined and radar-based products are truncated to about 10 weeks because of the SMAP radar failure). Upon completing the intensive CalVal phase of the mission the SMAP project will continue to enhance the products in the primary and extended geographic domains, in co-operation with the CalVal Partners, by continuing the comparisons over the existing core validation sites and inclusion of candidate sites that can address shortcomings.

SMAP↗
NASA NTRS2024

The Radio Fundamental Catalogue. I. Astrometry

We present the all-sky catalogue of absolute positions and estimates of correlated flux density of 21,942 compact radio sources determined from processing interferometric visibility data of virtually all VLBI observing sessions at 2 to 23~GHz from 72 programs suitable for absolute astrometry collected for 30 years. We used a novel technique of generation of a dataset of fused observables that allowed us to incorporate all available data in our analysis. The catalogue is the most complete and most precise to date. It forms the foundation and reference for positional astronomy, space geodesy, space navigation, population analysis of active galactic nuclei (AGNs), and provides calibrators for phase-referencing for differential astrometry and VLBI astrophysical observations. Its accuracy was evaluated through a detailed accounting systematic errors, rigorous decimation tests, comparison of different datasets, and comparison with other catalogues. The catalogue samples preferentially the AGNs with a strong contemporary parsec-scale synchrotron emission. Its milliarcsecond level positional accuracy allows association of these AGNs with detections in a wide range of the electromagnetic spectrum from low-frequency radio to gamma-rays and high energy neutrinos. We describe the innovative data processing and calibration technique in full detail, report the in depth analysis of random and systematic positional errors, and provide the list of associations with large surveys at different wavelengths.

Astronomy↗
NASA NTRS2012

Simulation Framework to Estimate the Performance of CO2 and O2 Sensing from Space and Airborne Platforms for the ASCENDS Mission Requirements Analysis

The Active Sensing of CO2 Emissions over Nights Days and Seasons (ASCENDS) mission recommended by the NRC Decadal Survey has a desired accuracy of 0.3% in carbon dioxide mixing ratio (XCO2) retrievals requiring careful selection and optimization of the instrument parameters. NASA Langley Research Center (LaRC) is investigating 1.57 micron carbon dioxide as well as the 1.26-1.27 micron oxygen bands for our proposed ASCENDS mission requirements investigation. Simulation studies are underway for these bands to select optimum instrument parameters. The simulations are based on a multi-wavelength lidar modeling framework being developed at NASA LaRC to predict the performance of CO2 and O2 sensing from space and airborne platforms. The modeling framework consists of a lidar simulation module and a line-by-line calculation component with interchangeable lineshape routines to test the performance of alternative lineshape models in the simulations. As an option the line-by-line radiative transfer model (LBLRTM) program may also be used for line-by-line calculations. The modeling framework is being used to perform error analysis, establish optimum measurement wavelengths as well as to identify the best lineshape models to be used in CO2 and O2 retrievals. Several additional programs for HITRAN database management and related simulations are planned to be included in the framework. The description of the modeling framework with selected results of the simulation studies for CO2 and O2 sensing is presented in this paper.

Plitau, Denis↗
DOE OSTI2026

A hybrid Penman-Monteith and machine learning model for simulating evapotranspiration and its components

Integrating physical processes with machine learning has advanced evapotranspiration (ET) simulation, yet most hybrid models fail to partition total ET into its components: soil evaporation (E) and vegetation transpiration (T). This study introduces Residual Neural Network–Penman–Monteith (RNN-PM), a novel hybrid dual-source ET model designed to overcome this limitation. The model synergizes the physically-based Penman–Monteith framework with three specialized residual neural networks trained to estimate key conductance parameters (canopy conductance, soil surface conductance, and aerodynamic conductance). Furthermore this explicit parameterization allows for the direct partitioning of total ET. Validation at National Ecological Observatory Network (NEON) flux sites using high-frequency partitioned E and T shows that RNN-PM reliably reproduces ET and the transpiration fraction (T/ET). For ET, the model achieves an average Kling–Gupta efficiency (KGE) of 0.89 and a root-mean-square error (RMSE) of 0.55 mm/day; for T/ET, the KGE is 0.87 with an RMSE of 0.06. Furthermore, RNN-PM demonstrates robust generalization, accurately simulating ET and its components well beyond the initial training dataset, even under extreme climatic conditions. This study extended the analysis by comparing the RNN-PM model with seven established dual-source ET models. The results indicate that RNN-PM outperforms both conventional machine learning models and purely physical process-based models in simulating ET components in most cases. Among the purely physical process-based dual-source models, those based on surface temperature decomposition showed improved performance as the leaf area index (LAI) decreased when evaluated against high-frequency ET component datasets. In contrast, the performance of conductance-based dual-source models declined with decreasing LAI. Although purely machine learning-based models can produce relatively accurate simulations of ET components, they often exhibit limited generalization capability, an issue that the RNN-PM model effectively overcomes. Ultimately, the RNN-PM model represents a significant advance in simulating ET components, offering a novel and scalable approach for improving the representation of land–atmosphere interactions in Earth system models.

54 ENVIRONMENTAL SCIENCES↗