Search NASA⌕ Search

SEARCH · Search NASA

Results for “performance proxy”

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 253 records · Page 14

Legacies of precipitation influence primary production in Panicum virgatum

Precipitation is a key driver of primary production worldwide, but primary production does not always track year-to-year variation in precipitation linearly. Instead, plant responses to changes in precipitation may exhibit time lags, or legacies of past precipitation. Legacies can be driven by multiple mechanisms, including persistent changes in plant physiological and morphological traits and changes to the physical environment, such as plant access to soil water. Here, we used three precipitation manipulation experiments in central Texas, USA to evaluate the magnitude, duration, and potential mechanisms driving precipitation legacies on aboveground primary production of the perennial C 4 grass, Panicum virgatum. Specifically, we performed a rainout shelter study, where eight genotypes grew under different precipitation regimes; a transplant study, where plants that had previously grown in a rainout shelter under different precipitation regimes were moved to a common environment; and a mesocosm study, where the effect of swapping precipitation regime was examined with a single genotype. Across these experiments, plants previously grown under wet conditions generally performed better than expected when exposed to drought. Panicum virgatum exhibited stronger productivity legacies of past wet years on current-year responses to drought than of past dry years on current-year responses to wet conditions. Additionally, previous year tiller counts, a proxy for meristem availability, were important in determining legacy effects on aboveground production. As climate changes and precipitation extremes—both dry and wet—become more common, these results suggest that populations of P. virgatum may become less resilient.

54 ENVIRONMENTAL SCIENCES↗

Gaining Perspective on Unconventional Well Design Choices through Play-level Application of Machine Learning Modeling

The recent development of unconventional oil and gas (O&G) reservoirs has led to an abundant hydrocarbon supply, both domestically and globally. However, there is a continued push to develop new and innovative approaches to improve exploration and extraction efficiencies and overall well productivity moving forward. Substantial improvements in unconventional O&G development are expected through optimized well completion and stimulation strategies aimed at maximizing well productivity. Optimizing well designs will require tailoring to the distinctive geologic conditions present for any newly placed well. To better evaluate the impact of well design attributes and their associated interactions on productivity in a major unconventional play, multivariate machine learning-based models that use empirical datasets were developed. A gradient boosted regression tree (GBRT) algorithm was applied. GBRT has been narrowly investigated for O&G applications but enables straightforward parametric importance and influence evaluation, as well as assessment of parameter interaction effects. Models were trained on well design and locational parameters that serve as a proxy for variable geologic conditions to estimate two types of productivity indicator response variables strongly correlated to estimated ultimate recovery (EUR). The dataset utilized consists of over 7,000 well observations that cover the majority of the productive region of the Marcellus Shale. Model performance was evaluated and algorithm parameters tuned by analyzing the goodness-of-fit for simulated results against observed data in a cross-validation approach. Models were found capable of 73–79 percent prediction accuracy on held out testing data of gas equivalent production and can be used to inform future well design and placement decisions for increasing EUR per well and improving overall field-level recovery. Study results indicate that Marcellus well performance improves most with upscaling perforated interval lengths and water and proppant volumes per foot; but relative productivity improvements are spatially dependent across the play. Finally, optimal combinations of water and proppant on well performance were found to vary depending on well location, emphasizing the utility of data-driven models capable of broad application across a play of interest for informing tailored well design approaches prior to their field deployment.

04 OIL SHALES AND TAR SANDS↗

The SRG/eROSITA All-Sky Survey: Exploring halo assembly bias with X-ray-selected superclusters

Numerical simulations indicate that the clustering of dark matter halos is not only dependent on the halo masses but has a secondary dependence on other properties, such as the assembly history of the halo. This phenomenon, known as the halo assembly bias (HAB), has been found mostly on galaxy scales; observational evidence on larger scales is scarce. In this work, we propose a novel method for exploring HAB on cluster scales using large samples of superclusters. Leveraging the largest-ever X-ray galaxy cluster and supercluster samples obtained from the first SRG/eROSITA all-sky survey, we constructed two subsamples of galaxy clusters that consist of supercluster members and isolated clusters, respectively. After correcting for the selection effects on redshift, mass, and survey depth, we computed the excess in the concentration of the intracluster gas of isolated clusters with respect to supercluster members, defined as δc gas ≡ c gas , ISO /c gas , SC –1 , to investigate the environmental effect on the concentration of clusters, a sign of HAB on cluster scales. We find that the average gas mass concentration of isolated clusters is a few percent higher than that of supercluster members, with a maximum significance of 2.8σ. The result for δc gas varies with the overdensity ratio, f, in supercluster identification, cluster mass proxies, and mass and redshift ranges but remains positive in almost all the measurements. We measure slightly larger δc gas when adopting a higher f for supercluster identification. The δc gas is also higher for low-mass and low-redshift clusters. We performed weak lensing analyses to compare the total mass concentration of the two classes and find a similar trend in total mass concentration as obtained from the gas mass concentration. Our results are consistent with the prediction of HAB on cluster scales, where halos located in denser environments are less concentrated; this trend is stronger for halos with lower masses and at lower redshifts. These phenomena can be explained by the fact that clusters in denser environments, such as superclusters, have experienced more mergers than isolated clusters in their assembling history. This work paves the way to explore HAB with X-ray superclusters and demonstrates that large samples of superclusters with X-ray and weak-lensing data can advance our understanding of the evolution of the large-scale structure.

79 ASTRONOMY AND ASTROPHYSICS↗

Cloud drop number concentrations over the western North Atlantic Ocean: seasonal cycle, aerosol interrelationships, and other influential factors

Cloud drop number concentrations (N d ) over the western North Atlantic Ocean (WNAO) are generally highest during the winter (DJF) and lowest in summer (JJA), in contrast to aerosol proxy variables (aerosoloptical depth, aerosol index, surface aerosol mass concentrations, surface cloud condensation nuclei (CCN) concentrations) that generally peak inspring (MAM) and JJA with minima in DJF. Using aircraft, satellite remote sensing, ground-based in situ measurement data, and reanalysis data, we characterize factors explaining the divergent seasonal cycles and furthermore probe into factors influencing N d on seasonal timescales. The results can be summarized well by features most pronounced in DJF, including features associated with cold-air outbreak (CAO) conditions such as enhanced values of CAO index, planetary boundary layer height (PBLH),low-level liquid cloud fraction, and cloud-top height, in addition to winds aligned with continental outflow. Data sorted into high- and low-N d days in each season, especially in DJF, revealed that all of these conditions were enhanced on the high-N d days, including reduced sea level pressure and stronger wind speeds. Although aerosols may be more abundant in MAM and JJA, the conditions needed to activate those particles into cloud droplets are weaker than in colder months, which is demonstrated by calculations of the strongest (weakest) aerosol indirect effects in DJF (JJA) based on comparing N d to perturbations in four different aerosol proxy variables (total and sulfate aerosol optical depth, aerosol index, surface mass concentration of sulfate). We used three machine learning models and up to 14 input variables to infer about most influential factors related to N d for DJF and JJA, with the best performance obtained with gradient-boosted regression tree (GBRT) analysis. The model results indicated that cloud fraction was the most important input variable, followed by some combination (depending on season) of CAO index and surface mass concentrations of sulfate and organic carbon. Future work is recommended to further understand aspects uncovered here such as impacts of free tropospheric aerosol entrainment on clouds, degree of boundary layer coupling, wet scavenging, and giant CCN effects on aerosol–N d relationships, updraft velocity, and vertical structure of cloud properties such as adiabaticity that impact the satellite estimation of N d .

54 ENVIRONMENTAL SCIENCES↗

Effect of Networking Density on the Patterning Performance of Molecular Layer Deposited Alucone Electron Beam/EUV Resists

One of the main barriers to continued device scaling in the era of extreme ultraviolet (EUV) lithography is the need for improved photoresist chemistries to address challenges such as poor EUV sensitivity, inadequate etch resistance, and pattern collapse. Metal-organic photoresists are a promising class of materials that can address many of these challenges, and among them, resists deposited via hybrid molecular layer deposition (MLD) have attracted interest for their unique advantages in thickness control, chemical homogeneity, and compatibility with vacuum processing. However, despite many successful demonstrations of patterning, little is known about how the molecular design of hybrid MLD resists affects their lithographic performance. In this work, we study the effect of the network structure, a common feature among all hybrid MLD resists, via a series of aluminum alkoxide ("alucone") negative tone resists with varying networking density. Their patterning mechanism is investigated via electron beam lithography (EBL)─a common proxy for EUV─and compared to their EUV-induced reactions studied via flood exposure and in situ characterization. In conclusion, we show that the resist with the least networking density demonstrates the best sensitivity and resolution, with the ability to resolve dense line/space gratings as small as 14 nm half pitch via EBL.

electron beam lithography↗

Discrete event performance prediction of speculatively parallel temperature-accelerated dynamics

Due to its unrivaled ability to predict the dynamical evolution of interacting atoms, molecular dynamics (MD) is a widely used computational method in theoretical chemistry, physics, biology, and engineering. Despite its success, MD is only capable of modeling time scales within several orders of magnitude of thermal vibrations, leaving out many important phenomena that occur at slower rates. The Temperature Accelerated Dynamics (TAD) method overcomes this limitation by thermally accelerating the state-to-state evolution captured by MD. Due to the algorithmically complex nature of the serial TAD procedure, implementations have yet to improve performance by parallelizing the concurrent exploration of multiple states. Here we utilize a discrete event-based application simulator to introduce and explore a new Speculatively Parallel TAD (SpecTAD) method. We investigate the SpecTAD algorithm, without a full-scale implementation, by constructing an application simulator proxy (SpecTADSim). Finally, following this method, we discover that a nontrivial relationship exists between the optimal SpecTAD parameter set and the number of CPU cores available at run-time. Furthermore, we find that a majority of the available SpecTAD boost can be achieved within an existing TAD application using relatively simple algorithm modifications.

97 MATHEMATICS AND COMPUTING↗

Hardware assisted fine-grained data movement

A processor includes a task scheduling unit and a compute unit coupled to the task scheduling unit. The task scheduling unit performs a task dependency assessment of a task dependency graph and task data requirements that correspond to each task of the plurality of tasks. Based on the task dependency assessment, the task scheduling unit schedules a first task of the plurality of tasks and a second proxy object of a plurality of proxy objects specified by the task data requirements such that a memory transfer of the second proxy object of the plurality of proxy objects occurs while the first task is being executed.

Hassaan, Muhammad Amber↗

Hardware assisted fine-grained data movement

A processor includes a task scheduling unit and a compute unit coupled to the task scheduling unit. The task scheduling unit performs a task dependency assessment of a task dependency graph and task data requirements that correspond to each task of the plurality of tasks. Based on the task dependency assessment, the task scheduling unit schedules a first task of the plurality of tasks and a second proxy object of a plurality of proxy objects specified by the task data requirements such that a memory transfer of the second proxy object of the plurality of proxy objects occurs while the first task is being executed.

Hassaan, Muhammad Amber↗

Homogenization of time series from Portugal and its former colonies for the period from the late 19th to the early 21st century

Here, three homogenization software packages – AnClim, RhTestsV3 and HOMER – have been applied to monthly and annual temperature and pressure series from Portugal and its ex-colonies in Africa and Asia to detect non-climatic breaks. Homogenization tests in both absolute and relative mode were performed with an a posteriori statistical assessment to select the most probable breakpoints. Time series from neighbouring stations as well as reanalyses data were used as reference series. The reanalyses data sets were used due to lack of suitable neighbouring stations and also to test reanalyses as proxy data. A list of probable shifts is produced and will be used to support future global reanalyses. Moreover, temperature and pressure series were adjusted using adjustment factors that were obtained by calculating the difference between certain periods before and after the breaks. The accuracy of the adjusted series was verified subjectively (visually) and objectively using three verification techniques: correlation coefficient, central root-mean-square error and standard deviation. A positive impact of the applied adjustment in terms of all three scores was reflected in a majority of cases.

abrupt breaks↗

Time-resolved analysis of Fermi gamma-ray bursts with fast- and slow-cooled synchrotron photon models

Time-resolved spectroscopy is performed on eight bright, long gamma-ray bursts (GRBs) dominated by single emission pulses that were observed with the Fermi Gamma-Ray Space Telescope. Fitting the prompt radiation of GRBs by empirical spectral forms such as the Band function leads to ambiguous conclusions about the physical model for the prompt radiation. Moreover, the Band function is often inadequate to fit the data. The GRB spectrum is therefore modeled with two emission components consisting of optically thin non-thermal synchrotron radiation from relativistic electrons and, when significant, thermal emission from a jet photosphere, which is represented by a blackbody spectrum. To produce an acceptable fit, the addition of a blackbody component is required in five out of the eight cases. We also find that the low-energy spectral index α is consistent with a synchrotron component with α = –0.81 ± 0.1. This value lies between the limiting values of α = –2/3 and α = –3/2 for electrons in the slow- and fast-cooling regimes, respectively, suggesting ongoing acceleration at the emission site. The blackbody component can be more significant when using a physical synchrotron model instead of the Band function, illustrating that the Band function does not serve as a good proxy for a non-thermal synchrotron emission component. The temperature and characteristic emission-region size of the blackbody component are found to, respectively, decrease and increase as power laws with time during the prompt phase. In addition, we find that the blackbody and non-thermal components have separate temporal behaviors as far as their respective flux and spectral evolutions.

79 ASTRONOMY AND ASTROPHYSICS↗

A family of independent Variable Eddington Factor methods with efficient preconditioned iterative solvers

We present a family of discretizations for the Variable Eddington Factor (VEF) equations that have high-order accuracy on curved meshes and efficient preconditioned iterative solvers. The VEF discretizations are combined with the Discontinuous Galerkin transport discretization from to form effective high-order, linear transport methods. The VEF discretizations are derived by extending the unified analysis of Discontinuous Galerkin methods for elliptic problems presented by Arnold et al. to the VEF equations. This framework is used to define analogs of the interior penalty, second method of Bassi and Rebay, minimal dissipation local Discontinuous Galerkin, and continuous finite element methods. The analysis of subspace correction preconditioners, which use a continuous operator to iteratively precondition the discontinuous discretization, is extended to the case of the non-symmetric VEF system. Numerical results demonstrate that the VEF discretizations have arbitrary-order accuracy on curved meshes, preserve the thick diffusion limit, and are effective on a proxy problem from thermal radiative transfer in both outer transport iterations and inner preconditioned linear solver iterations. We demonstrate that the VEF solution converges to the S N transport solution as the mesh is refined on both problems with smooth and non-smooth behavior in angle. Parallel performance studies show that the interior penalty VEF discretization's linear solve weak scales out to 1024 processors and strong scales well on a single node. Particular attention is paid to the parallel performance of the VEF algorithm when used in combination with a parallel block Jacobi transport sweep.

97 MATHEMATICS AND COMPUTING↗

Thread-level parallelization and optimization of NWChem for the Intel MIC architecture

In the multicore era it was possible to exploit the increase in on-chip parallelism by simply running multiple MPI processes per chip. Unfortunately, manycore processors' greatly increased thread- and data-level parallelism coupled with a reduced memory capacity demand an altogether different approach. In this paper we explore augmenting two NWChem modules, triples correction of the CCSD(T) and Fock matrix construction, with OpenMP in order that they might run efficiently on future manycore architectures. As the next NERSC machine will be a self-hosted Intel MIC (Xeon Phi) based supercomputer, we leverage an existing MIC testbed at NERSC to evaluate our experiments. In order to proxy the fact that future MIC machines will not have a host processor, we run all of our experiments in native mode. We found that while straightforward application of OpenMP to the deep loop nests associated with the tensor contractions of CCSD(T) was sufficient in attaining high performance, significant e ort was required to safely and efeciently thread the TEXAS integral package when constructing the Fock matrix. Ultimately, our new MPI+OpenMP hybrid implementations attain up to 65× better performance for the triples part of the CCSD(T) due in large part to the fact that the limited on-card memory limits the existing MPI implementation to a single process per card. Additionally, we obtain up to 1.6× better performance on Fock matrix constructions when compared with the best MPI implementations running multiple processes per card.

CCSD(T)↗

Data from A Mass Spectrometry-Based High-Throughput Screening Method for Engineering Fatty Acid Synthases with Improved Production of Medium Chain Fatty Acids

Microbial cell factories have been extensively engineered to produce free fatty acids (FFAs) as key components of crucial nutrients, soaps, industrial chemicals, and fuels. However, our ability to control the composition of microbially synthesized FFAs is still limited, particularly, for producing medium‐chain fatty acids (MCFAs). This is mainly due to the lack of high‐throughput approaches for FFA analysis to engineer enzymes with desirable product specificity. Here we report a mass spectrometry (MS)‐based method for rapid profiling of MCFAs in Saccharomyces cerevisiae by using membrane lipids as a proxy. In particular, matrix‐assisted laser desorption/ionization time‐of‐flight (MALDI‐ToF) MS was used to detect shorter acyl chain phosphatidylcholines from membrane lipids and a higher m/z peak ratio at 730 and 758 was used as an indication for improved MCFA production. This colony‐based method can be performed at a rate of ~2 s per sample, representing a substantial improvement over gas chromatography‐MS (typically >30 min per sample) as the gold standard method for FFA detection. To demonstrate the power of this method, we performed site‐saturation mutagenesis of the yeast fatty acid synthase and identified nine missense mutations that resulted in improved MCFA production relative to the wild‐type strain. Colony‐based MALDI‐ToF MS screening provides an effective approach for engineering microbial fatty acid compositions in a high‐throughput manner.

Conversion↗

Thread-Level Parallelization and Optimization of NWChem for the Intel MIC Architecture

In the multicore era it was possible to exploit the increase in on-chip parallelism by simply running multiple MPI processes per chip. Unfortunately, manycore processors' greatly increased thread- and data-level parallelism coupled with a reduced memory capacity demand an altogether different approach. In this paper we explore augmenting two NWChem modules, triples correction of the CCSD(T) and Fock matrix construction, with OpenMP in order that they might run efficiently on future manycore architectures. As the next NERSC machine will be a self-hosted Intel MIC (Xeon Phi) based supercomputer, we leverage an existing MIC testbed at NERSC to evaluate our experiments. In order to proxy the fact that future MIC machines will not have a host processor, we run all of our experiments in tt native mode. We found that while straightforward application of OpenMP to the deep loop nests associated with the tensor contractions of CCSD(T) was sufficient in attaining high performance, significant effort was required to safely and efficiently thread the TEXAS integral package when constructing the Fock matrix. Ultimately, our new MPI OpenMP hybrid implementations attain up to 65x better performance for the triples part of the CCSD(T) due in large part to the fact that the limited on-card memory limits the existing MPI implementation to a single process per card. Additionally, we obtain up to 1.6x better performance on Fock matrix constructions when compared with the best MPI implementations running multiple processes per card.

97 MATHEMATICS AND COMPUTING↗

Development of lean, efficient, and fast physics-framed deep-learning-based proxy models for subsurface carbon storage

In this work, we present deep-learning-based surrogate models for CCUS developed with four different algorithms and a physics-framed two-phase flow problem involving displacement of water by CO 2 . The deep-learning models were trained using 3D datasets describing the pressure plume, CO 2 saturation plume, and water extraction rate generated by numerical simulation. The hyperparameters defining the architecture of the neural networks were optimized to determine the slimmest network size and training parameters that give the most efficient performance at the least training cost. To develop a robust model that closely mimics the governing physical laws, the discretized form of the two-phase fluid transport equation was used to formulate the supervised deep-learning task. The algorithms investigated in this study predicted the data to above 95% accuracy, with the multi-layer perceptron model demonstrating the best performance by balancing training speed, prediction time, and prediction accuracy with lean network capacity. Furthermore, the surrogate models simultaneously predict reservoir pressure and CO 2 saturation in every grid block, including the surface well extraction rate and bottomhole pressure, at all simulation times for a given static model realization in just a few seconds on a standard desktop computer. A key outcome of this study is that limits can be placed on network design parameters to avoid over designing neural networks, with associated efficiencies in training and prediction times. This is very useful because large volumes of data may be generated in CCUS projects and over-design of neural network architectures imposes penalties that are antithetical to the goal of near-real time forecasting.

58 GEOSCIENCES↗

X-ray absorption spectroscopy and theoretical investigations of the effect of extended ligands in potassium organic matter interaction

Potassium (K) is an essential nutrient for plant growth, and despite its abundance in soil, most of the K is structurally bound in minerals, limiting its bioavailability and making this soil K reservoir largely inaccessible to plants. Microbial biochemical weathering has been shown to be a promising pathway to sustainably increase plant available K. However, the mechanisms underpinning microbial K uptake, transformation, storage, and sharing are poorly resolved. Here, to better understand the controls on microbial K transformations, we performed K K-edge x-ray absorption near-edge structure (XANES) spectroscopy on K-organic salts, including acetate, citrate, nitrate, oxalate, and tartrate, which are frequently observed as low molecular weight organic acids secreted by soil microbes, as well as humic acid, which acts as a proxy for higher molecular weight organic acids. The organic salts display feature-rich K XANES spectra, each demonstrating numerous unique features spanning ~13 eV range across the absorption edge. In contrast, the spectra for humic acid have one broad, wide feature across the same energy range. We used a combination of time-dependent density functional theory and the Bethe–Salpeter equation based approach within the OCEAN code to simulate the experimental spectra for K-nitrate (KNO 3 ) and K-citrate [K 3 (C 6 H 5 O 7 )·H 2 O] to identify the electronic transitions that give rise to some of the outlying and unique spectral features in the organic salts. KNO 3 has both the lowest and highest lying energy features, and K 3 (C 6 H 5 O 7 )·H 2 O is produced by several soil microbes and is effective at mineral weathering. Our results analyze the K-organic salt bonding in detail to elucidate why the spectral shapes differ and indicate that the K K-edge XANES spectra are associated with the entire ligand despite similar first-shell bonding environments around the K center. The improved understanding of K bonding environments with organic ligands and their use for interpretation of the K-XANES spectra provides an important toolkit to understand how K is transformed by microbial processes and made bioavailable for plant uptake.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Effect of the space weather conditions on the Earth magnetosphere soft X-ray emissivity

The aim of the study is to model and characterize the soft X-ray emissivity on the Earth magnetosphere for different space weather conditions (SWC), providing information to interpret the soft X-ray measurements of the Solar wind Magnetosphere Ionosphere Link Explorer space mission. The MHD code pluto in spherical coordinates is used to perform parametric studies with respect to the solar wind (SW) dynamic pressure (considering density and velocity effects independently) as well as the IMF intensity and orientation, predicting the soft X-ray emissivity for different SWC. The integrated soft X-ray emissivity inside the magnetosheath is calculated as a proxy of the soft X-ray emission dependencies with the SWC independently of the satellite orbit and camera line of sight. The analysis indicates fluctuations of the interplanetary magnetic field (IMF) orientation and magnitude may significantly affect the measured soft X-ray emission although changes in the SW dynamic pressure should be the main source of variability. The southward IMF orientation leads to the configuration with the largest soft X-ray emissivity and northward to the lowest. Strongly distorted magnetospheres explored in configurations showing SW and IMF parameters comparable to the impact of interplanetary coronal mass ejections may show a decrease of the soft X-ray emissivity as the IMF magnitude increases, explained by the strong magnetosphere compression and constriction of the magnetosheath region where the soft X-ray emissivity maximum is located. The simulations also indicate large excursions of the soft X-ray emissivity maximum inside the magnetosheath as the IMF magnitude and SW dynamic pressure fluctuate particularly for radial and ecliptic IMF orientations.

Earth↗

Using automated machine learning for the upscaling of gross primary productivity

Estimating gross primary productivity (GPP) over space and time is fundamental for understanding the response of the terrestrial biosphere to climate change. Eddy covariance flux towers provide in situ estimates of GPP at the ecosystem scale, but their sparse geographical distribution limits larger-scale inference. Machine learning (ML) techniques have been used to address this problem by extrapolating local GPP measurements over space using satellite remote sensing data. However, the accuracy of the regression model can be affected by uncertainties introduced by model selection, parameterization, and choice of explanatory features, among others. Recent advances in automated ML (AutoML) provide a novel automated way to select and synthesize different ML models. In this work, we explore the potential of AutoML by training three major AutoML frameworks on eddy covariance measurements of GPP at 243 globally distributed sites. We compared their ability to predict GPP and its spatial and temporal variability based on different sets of remote sensing explanatory variables. Explanatory variables from only Moderate Resolution Imaging Spectroradiometer (MODIS) surface reflectance data and photosynthetically active radiation explained over 70 % of the monthly variability in GPP, while satellite-derived proxies for canopy structure, photosynthetic activity, environmental stressors, and meteorological variables from reanalysis (ERA5-Land) further improved the frameworks' predictive ability. We found that the AutoML framework Auto-sklearn consistently outperformed other AutoML frameworks as well as a classical random forest regressor in predicting GPP but with small performance differences, reaching an r 2 of up to 0.75. We deployed the best-performing framework to generate global wall-to-wall maps highlighting GPP patterns in good agreement with satellite-derived reference data. This research benchmarks the application of AutoML in GPP estimation and assesses its potential and limitations in quantifying global photosynthetic activity.

54 ENVIRONMENTAL SCIENCES↗