Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Statistical accuracy”

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 1,153 records · Page 64
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Database and deep-learning scalability of anharmonic phonon properties by automated brute-force first-principles calculations

Understanding the anharmonic phonon properties of crystal compounds—such as phonon lifetimes and thermal conductivities—is essential for investigating and optimizing their thermal transport behaviors. These properties also impact optical, electronic, and magnetic characteristics through interactions between phonons and other quasiparticles and fields. In this study, we develop an automated first-principles workflow to calculate anharmonic phonon properties and build a comprehensive database encompassing more than 6500 inorganic compounds. Utilizing this dataset, we train a graph neural network model to predict thermal conductivity values and spectra from structural parameters, demonstrating a scaling law in which prediction accuracy improves with increasing training data size. High-throughput screening with the model enables the identification of materials exhibiting extreme thermal conductivities—both high and low. The resulting database offers valuable insights into the anharmonic behavior of phonons, thereby accelerating the design and development of advanced functional materials.

Ohnishi, Masato [University of Tokyo (Japan); Inst↗
DOE OSTI2022

How Many Elements Matter?

Some studies of stars' multielement abundance distributions suggest at least 5–7 significant dimensions, but others show that many elemental abundances can be predicted to high accuracy from [Fe/H] and [Mg/Fe] (or [Fe/H] and age) alone. We show that both propositions can be, and are, simultaneously true. We adopt a machine-learning technique known as normalizing flow to reconstruct the probability distribution of Milky Way disk stars in the space of 15 elemental abundances measured by APOGEE. Conditioning on T eff and $\mathrm{log}\,g$ minimizes the differential systematics. After further conditioning on [Fe/H] and [Mg/Fe], the residual scatter for most abundances is σ [X/H] ≲ 0.02 dex, consistent with APOGEE's reported statistical uncertainties of ~0.01–0.015 dex and intrinsic scatter of 0.01–0.02 dex. Despite the small scatter, residual abundances display clear correlations between elements, which we show are too large to be explained by measurement uncertainties or by the finite sampling noise. We must condition on at least seven elements to reduce the correlations to a level consistent with the observational uncertainties. Our results demonstrate that cross-element correlations are a much more sensitive probe of a hidden structure than dispersion, and they can be measured precisely in a large sample even if the star-by-star measurement noise is comparable to the intrinsic scatter. We conclude that many elements have an independent story to tell, even for the mundane disk stars and elements produced by the core-collapse and Type Ia supernovae. The only way to learn these lessons is to measure the abundances directly, and not merely infer them.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Measurements of the inclusive W and Z boson production cross sections and their ratios in proton-proton collisions at $\sqrt{s}=13.6$ TeV

Measurements are presented of the W and Z boson production cross sections in proton-proton collisions at a center-of-mass energy of 13.6 TeV. Data collected in 2022 and corresponding to an integrated luminosity of 5.01 fb −1 with one or two identified muons in the final state are analyzed. The results for the products of total inclusive cross sections and branching fractions for muonic decays of W and Z bosons are 11.93 ± 0.08 (syst) ± 0.17 (lumi) $^{+0.07}_{−0.07}$ (acceptance) nb for W + boson production, 8.86 ± 0.06 (syst) ± 0.12 (lumi) $^{+0.05}_{−0.06}$ (acceptance) nb for W − boson production, and 2.021 ± 0.009 (syst) ± 0.028 (lumi) $^{+0.011}_{−0.013}$ (acceptance) nb for the Z boson production in the dimuon mass range of 60–120 GeV, all with negligible statistical uncertainties. Furthermore, the corresponding fiducial cross sections, as well as cross section ratios for both fiducial and total phase space, are provided. The ratios include charge-separated results for W boson production (W + and W − ) and the sum of the two contributions (W ± ), each relative to the measured Z boson production cross section. Additionally, the ratio of the measured cross sections for W + and W − boson production is reported. All measurements are in agreement with theoretical predictions, calculated at next-to-next-to-leading order accuracy in quantum chromodynamics.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

Large eddy simulation of atmospheric boundary layer flow over complex terrain in comparison with RANS simulation and on-site measurements under neutral stability condition

Large eddy simulation (LES) of the atmospheric boundary layer (ABL) flow over complex terrain is presented with a validation using meteorological tower (met-tower) data through an improved neutral stability sampling approach. The proposed stability sampling procedure includes a condition based on the most-likely occurrence time-periods of the neutral ABL and reduces the variabilities of the conditional wind statistics calculated at the met-towers in comparison to our previous work. The ABL flow simulations are carried out over a potential wind site with a prominent hill based using the OpenFOAM-based simulator for on/off-shore wind farm applications by applying the Lagrangian-averaged scale-invariant dynamic sub-grid scale turbulence model. A low-dissipative scale-selective discretization scheme for the non-linear convection term in the LES governing equation is adopted implicitly to ensure both the second-order accuracy and bounded solution. The LES inflow is generated through a precursor method with a “tiling” approach based on the flow driving parameters obtained from a corresponding Reynolds-averaged Navier–Stokes (RANS) simulation. Overall, the averaged wind velocity profiles predicted by the LES approach at all met-tower locations show a similar tendency as the RANS results, which are also in reasonable agreement with the met-tower data. An obvious difference in wind speed standard deviation profiles is seen between LES and RANS, especially at regions downstream of the hill edge, where the LES shows under-predicted results at the highest measurement levels in comparison to the tower data. The computational costs of the LES are found to be about 20 times higher than the RANS simulations.

Energy & Fuels↗
DOE OSTI2023

Better Than Worst-Case Decoding for Quantum Error Correction

The overheads of classical decoding for quantum error correction in cryogenic quantum systems grow rapidly with the number of logical qubits and their correction code distance. Decoding at room temperature is bottlenecked by refrigerator I/O bandwidth while cryogenic on-chip decoding is limited by area/power/thermal budget. To overcome these overheads, we are motivated by the observation that in the common case (over 90% of the time), error correction 'syndromes' are fairly trivial with high redundancy / sparsity, since the error correction codes are over-provisioned to be able to correct for uncommon worst-case complex scenarios (to ensure substantially low logical error rates). If suitably exploited, these trivial scenarios can be handled with insignificant overhead, thereby alleviating any bottlenecks towards handling the worst-case scenarios by state-of-the-art means. Here, we propose Better Than Worst-Case Decoding for Quantum Error Correction, targeting cryogenic quantum systems and Surface Code, consisting of: On-chip Clique Decoder: An extremely lightweight decoder for correcting trivial common-case errors, designed for the cryogenic domain. The decoder is implemented and evaluated for SFQ logic. Statistical Off-chip Bandwidth Allocation: A statistical confidence-based technique for allocation of off-chip decoding bandwidth, to efficiently handle the rare complex decodes that are not covered by the Clique Decoder. Decode-Overflow Execution Stalling: A method to stall circuit execution, for the worst-case scenarios in which the provisioned off-chip bandwidth is insufficient to complete all requested off-chip decodes. In all, BTWC decoding achieves 70-99+% off-chip bandwidth elimination across a range of logical and physical error rates, without significantly sacrificing the accuracy of a state-of-the-art off-chip decoder. Further, it achieves 10-1000x bandwidth reduction over prior bandwidth reduction techniques, as well as 15-37x resource overhead reduction compared to prior on-chip decoding.

cryogenic optimization↗
DOE OSTI2022

Sustainability Trait Modeling of Field-Grown Switchgrass (Panicum virgatum) Using UAV-Based Imagery

Unmanned aerial vehicles (UAVs) provide an intermediate scale of spatial and spectral data collection that yields increased accuracy and consistency in data collection for morphological and physiological traits than satellites and expanded flexibility and high-throughput compared to ground-based data collection. In this project, we used UAV-based multispectral imagery collected from MicaSense RedEdge-M on a DJI Matrice 600 Pro for automated phenotyping of field-grown switchgrass (Panicum virgatum), a leading bioenergy feedstock. The raw images were processed with Pix4D Mapper to create the reflectance data, and vegetation indices were calculated from a UAV-based multispectral camera. Statistical models were developed for rust disease caused by Puccinia novopanici, leaf chlorophyll, nitrogen, and lignin contents. For the first time, UAV remote sensing technology was used to explore the potential for multiple traits associated with sustainable production of switchgrass. One statistical model was developed for each individual trait based on the statistical correlation between vegetation indices and the corresponding trait.

CBI↗
NASA NTRS2008

Assessment of 3D Codes for Predicting Liner Attenuation in Flow Ducts

This paper presents comparisons of seven propagation codes for predicting liner attenuation in ducts with flow. The selected codes span the spectrum of methods available (finite element, parabolic approximation, and pseudo-time domain) and are collectively representative of the state-of-art in the liner industry. These codes are included because they have two-dimensional and three-dimensional versions and can be exported to NASA's Columbia Supercomputer. The basic assumptions, governing differential equations, boundary conditions, and numerical methods underlying each code are briefly reviewed and an assessment is performed based on two predefined metrics. The two metrics used in the assessment are the accuracy of the predicted attenuation and the amount of wall clock time to predict the attenuation. The assessment is performed over a range of frequencies, mean flow rates, and grazing flow liner impedances commonly used in the liner industry. The primary conclusions of the study are (1) predicted attenuations are in good agreement for rigid wall ducts, (2) the majority of codes compare well to each other and to approximate results from mode theory for soft wall ducts, (3) most codes compare well to measured data on a statistical basis, (4) only the finite element codes with cubic Hermite polynomials capture extremely large attenuations, and (5) wall clock time increases by an order of magnitude or more are observed for a three-dimensional code relative to the corresponding two-dimensional version of the same code.

Watson, W. R.↗
NASA NTRS2019

Lightning Strike Distance Distribution Beyond a Preexisting Lightning Area

The 45th Weather Squadron (45 WS) asked the Applied Meteorology Unit (AMU) to review the 30-year-old, lightning stand-off distances of 5 nautical miles (nmi) for applicability to today's operations. This was based on the realization that previous lightning strike distance studies did not match how 45 WS issues lightning warnings (Roeder, 2008). The previous lightning distance studies were from the point of origin of the lightning or the average starting location that would tend to be in the core of the thunderstorm. However, the 45 WS issues lightning warnings based on the edge of a preexisting lightning area. Before beginning the AMU project, it took several years to develop a method to calculate a distance distribution beyond a preexisting area (Roeder, 2015). The AMU pulled Lightning Detection and Ranging (LDAR) sensor data from 1/1/2013 to 12/31/2013. This dataset consisted of 37 million individual source data points from the LDAR sensors. Only sources within 50 km north, south, east or west of the LDAR grid center were included in the dataset. This limited the use of LDAR data to that with the greatest accuracy of source detection and increased data processing. Points were grouped into flashes based on spatial and temporal criteria. Based on the sensitivity analysis the AMU performed on the flash clustering algorithm, a time value of 0.3 seconds was found to model flashes adequately. Distance parameters were tested from 1,500 to 7,500 meters (m) in 500 m increments. Distance parameters of both 3,000 m and 4,000 m produced results in the plotting tool that were most representative of the physical behavior of lightning. Thus statistics were gathered for the most representative of these spatial and temporal criteria on the flash size and the polygon expansion distance in order to find the correct data distributions. The best fit curves for the LDAR polygon expansion frequency vs. distances for both the 3 kilometer (km) and 4 km distance threshold values were exponential decay functions and had R2 values of > 0.998, indicating good model fits. The equations of the best fit curves were then used to calculate a desired safety radius of 4 nmi for either 3 km or 4 km distance threshold criteria. The AMU analysis concludes the safe reduction of the 5 nmi lightning warning circles to 4 nmi should improve the operational impact by 36% if based on distance from the center of the property area being protected. If based on the edge of the property being protected, then the reduction is 4.5 nmi to 4 nmi and the operational impact is decreased by 21%. For the 6 nmi lightning warning circles, the recommended 4 nmi stand-off distance will result in a safe reduction of operational impact of 31% if based on the center of the area being protected, or 16% if based on the edge of the property.

Flash clustering algorithm↗
DOE OSTI2020

Unrecognized Sources of Uncertainties ( USU ) in Experimental Nuclear Data

Evaluated nuclear data uncertainties reported in the literature or archived in data libraries are often perceived as unrealistic, most often because they are thought to be too small. The impact of this issue in applied nuclear science has been discussed widely in recent years. Commonly suggested causes are: poor estimates of specific error components, neglect of uncertainty correlations, and overlooked known error sources. However, instances have been reported where very careful, objective assessments of all known error sources have been made with realistic error magnitudes and correlations provided, yet the resulting evaluated uncertainties still appear to be inconsistent with observed scatter of predicted mean values. These discrepancies might be attributed to significant unrecognized sources of uncertainty ( USU ) that limit the accuracy to which these physical quantities can be determined. The objective of our work reported herein has been to develop qualitative and quantitative procedures for revealing and including USU estimates in nuclear data evaluations involving experimental input data. This paper identifies several specific clues that can be explored by evaluators in identifying the existence of USU . It then describes numerical procedures we have introduced to generate quantitative estimates of USU magnitudes. Key requirements for these procedures to be viable are that sufficient numbers of data points be available, for statistical reasons, and that additional supporting information about the measurements be provided by the experimenters. Several realistic examples are described here to illustrate these procedures and demonstrate their outcomes and limitations. Our work strongly supports the view that USU is an important issue in nuclear data evaluation, with significant consequences for applications, and that this topic warrants further investigation by the nuclear science community.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Stream Temperature Prediction in a Shifting Environment: Explaining the Influence of Deep Learning Architecture

Stream temperature is a fundamental control on ecosystem health. Recent efforts incorporating process guidance into deep learning models for predicting stream temperature have been shown to outperform existing statistical and physical models. This performance is in part because deep learning architectures can actively learn spatiotemporal relationships that govern how water and energy propagate through a river network. However, exploration of how spatiotemporal awareness and process guidance influence a model's generalizability under shifting environmental conditions such as climate change is limited. Here, we use Explainable Artificial Intelligence (XAI) to interrogate how differing deep learning architectures affect a model's learned spatial and temporal dependencies, and how those learned dependencies affect a model's ability to maintain high accuracy when applied to unseen environmental conditions. Using the Delaware River Basin in the northeastern United States as a test case, we compare two spatiotemporally aware process–guided deep learning models for predicting stream temperature (a recurrent graph convolution network—RGCN, and a temporal convolution graph model—Graph WaveNet). Both models achieve equally high predictive performance when testing data are well represented in the training data (test root mean squared errors of 1.64°C and 1.65°C); however, Graph WaveNet significantly outperforms RGCN in 4 out of 5 experiments where test partitions represent different types of unseen environmental conditions. XAI results show that the architecture of Graph WaveNet leads to learned spatial relationships with greater fidelity to physical processes, and that this fidelity improves the generalizability of the model when applied to shifting and/or unseen environmental conditions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Point Adversarial Self-Mining: A Simple Method for Facial Expression Recognition

In this article, we propose a simple yet effective approach, called point adversarial self mining (PASM), to improve the recognition accuracy in facial expression recognition (FER). Unlike previous works focusing on designing specific architectures or loss functions to solve this problem, PASM boosts the network capability by simulating human learning processes: providing updated learning materials and guidance from more capable teachers. Specifically, to generate new learning materials, PASM leverages a point adversarial attack method and a trained teacher network to locate the most informative position related to the target task, generating harder learning samples to refine the network. The searched position is highly adaptive since it considers both the statistical information of each sample and the teacher network capability. Other than being provided new learning materials, the student network also receives guidance from the teacher network. After the student network finishes training, the student network changes its role and acts as a teacher, generating new learning materials and providing stronger guidance to train a better student network. Here, the adaptive learning materials generation and teacher/student update can be conducted more than one time, improving the network capability iteratively. Extensive experimental results validate the efficacy of our method over the existing state of the arts for FER.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Status and prospects of Muon g-2 experiment

This article reviews the muon g-2 experiment, a cornerstone in precision tests of the Standard Model of particle physics. The experiment measures the anomalous magnetic moment of the muon with unprecedented accuracy, seeking potential discrepancies between theoretical predictions and experimental results that might indicate physics beyond the Standard Model. We trace the evolution of this measurement from its beginnings at CERN in the 1960s to the current state-of-the-art experiment at Fermilab, highlighting the remarkable engineering achievements required to achieve parts-per-billion precision. Recent results from Runs 1-3 have achieved a systematic uncertainty of 70 ppb, exceeding design goals, while ongoing theoretical calculations continue to refine predictions. Despite these advances, the analysis remains statistics-limited, with continued data collection and novel experimental approaches at MUonE and J-PARC promising further insights into this fundamental physical quantity.

Karuza, Marin [INFN, Trieste; Rijeka U.] (ORCID:00↗
NASA NTRS1996

Land Surface Temperature Measurements form EOS MODIS Data

We have developed a physics-based land-surface temperature (LST) algorithm for simultaneously retrieving surface band-averaged emissivities and temperatures from day/night pairs of MODIS (Moderate Resolution Imaging Spectroradiometer) data in seven thermal infrared bands. The set of 14 nonlinear equations in the algorithm is solved with the statistical regression method and the least-squares fit method. This new LST algorithm was tested with simulated MODIS data for 80 sets of band-averaged emissivities calculated from published spectral data of terrestrial materials in wide ranges of atmospheric and surface temperature conditions. Comprehensive sensitivity and error analysis has been made to evaluate the performance of the new LST algorithm and its dependence on variations in surface emissivity and temperature, upon atmospheric conditions, as well as the noise-equivalent temperature difference (NE(Delta)T) and calibration accuracy specifications of the MODIS instrument. In cases with a systematic calibration error of 0.5%, the standard deviations of errors in retrieved surface daytime and nighttime temperatures fall between 0.4-0.5 K over a wide range of surface temperatures for mid-latitude summer conditions. The standard deviations of errors in retrieved emissivities in bands 31 and 32 (in the 10-12.5 micrometer IR spectral window region) are 0.009, and the maximum error in retrieved LST values falls between 2-3 K. Several issues related to the day/night LST algorithm (uncertainties in the day/night registration and in surface emissivity changes caused by dew occurrence, and the cloud cover) have been investigated. The LST algorithms have been validated with MODIS Airborne Simulator (MAS) dada and ground-based measurement data in two field campaigns conducted in Railroad Valley playa, NV in 1995 and 1996. The MODIS LST version 1 software has been delivered.

Wan, Zhengming↗
NASA NTRS2018

Assessing Version 4 of the SMAP L4_SM Data Product

Version 4 of the SMAP Level 4 Surface and Root Zone Soil Moisture (L4_SM) product benefits from an improved land surface modeling system. Surface soil moisture is typically drier by several volumetric percent in Version 4 compared to Version 3, whereas root zone soil moisture is wetter in Version 4 in some regions and drier in others. Results from core validation site comparisons show that Version 4 of the L4_SM data product meets the accuracy requirement, which is formulated in terms of the RMSE after removal of the long-term mean difference (ubRMSE). The overall ubRMSE of the 3-hourly L4_SM data at the 9 km scale is 0.039 m3 m-3 for surface soil moisture and 0.029 m3 m-3 for root zone soil moisture, below the 0.04 m3 m-3 requirement. L4_SM surface and root zone soil moisture estimates are more skillful than model-only simulation estimates that are not informed by SMAP brightness temperature observations, with statistically significant improvements at the 5% level for surface soil moisture R and anomaly R values. Results from comparisons of the L4_SM product to in situ measurements from more than 400 sparse network sites corroborate the core validation site results.

Reichle, R.↗
DOE OSTI2026

SALSAA: a statistical approach to line shapes from an average atom

Ion-Stark line broadening is a key density diagnostic for hot dense plasmas relevant to inertial fusion and astrophysics. It is caused by interactions of a radiating ion with nearby perturbing ions, whose electric microfields lead to changes in bound-bound transition energies. Ion-Stark broadening becomes increasingly difficult to compute for complex, many-electron ions with myriad transitions. In this paper, we propose a simplified approach to ion-Stark broadening based on self-consistent ion distributions and electronic structure from an average-atom model. We find that this approach reproduces the line shape predictions of one traditional method for high-n K-shell emission lines in aluminum ions with accuracy sufficient for density diagnostics in thermal plasmas with equal ion and electron temperatures. We expect that this approach can be extended to provide a reasonable picture of ion-Stark broadening in many-electron ions, enabling rapid calculations of line profiles in complex spectra.

average-atom↗
DOE OSTI2025

ARCH: Large-scale knowledge graph via aggregated narrative codified health records analysis

Objective: Electronic health record (EHR) systems contain a wealth of clinical data stored as both codified data and free-text narrative notes (NLP). The complexity of EHR presents challenges in feature representation, information extraction, and uncertainty quantification. Here, to address these challenges, we proposed an efficient Aggregated naRrative Codified Health (ARCH) records analysis to generate a large-scale knowledge graph (KG) for a comprehensive set of EHR codified and narrative features. Methods: Using data from 12.5 million Veterans Affairs patients, ARCH first derives embedding vectors and generates similarities along with associated p-values to measure the strength of relatedness between clinical features with statistical certainty quantification. Next, ARCH performs a sparse embedding regression to remove indirect linkage between features to build a sparse KG. Finally, ARCH was validated on various clinical tasks, including detecting known relationships between entity pairs, predicting drug side effects, disease phenotyping, as well as sub-typing Alzheimer’s disease patients. Results: ARCH produces high-quality clinical embeddings and KG for over 60,000 codified and narrative EHR concepts. The KG and embeddings are visualized in the R-shiny powered web-API.3 ARCH achieved high accuracy in detecting EHR concept relationships, with AUCs of 0.926 (codified) and 0.861 (NLP) for similar EHR concepts, and 0.810 (codified) and 0.843 (NLP) for related pairs. It detected drug side effects with a 0.723 AUC, which improved to 0.826 after fine-tuning. Using both codified and NLP features, the detection power increased significantly. Compared to other methods, ARCH has superior accuracy and enhances weakly supervised phenotyping algorithms’ performance. Notably, it successfully categorized Alzheimer’s patients into two subgroups with varying mortality rates. Conclusion: The proposed ARCH algorithm generates large-scale high-quality semantic representations and knowledge graph for both codified and NLP EHR features, useful for a wide range of predictive modeling tasks.

Electronic health records↗
NASA NTRS1986

Optimum retrieval of precipitable water fields from VAS and surface data

The use of VAS radiances and conventional surface data to determine the optimum resolution and accuracy of low-level precipitable water fields retrieved from geosynchronous satellite observations is examined. VAS retrievals of precipitated water (PW) are compared with various channel selections in the regression algorithm and with the split window algorithm. The split window algorithm and the linear regression algorithm are described. Consideration is given to the sounding-field-of-view (SFOV) resolution; the data reveal that the optimum resolution for the split window algorithm is at 15-60 km resolution, and 30-60 km SFOV resolution is optimal for the regression method. Statistics and images from PW retrieval experiments conducted on July 13, 1981 are presented and utilized to determine optimum channel selection. A fast contouring method for VAS sounding images is proposed.

Chesters, Dennis↗
NASA NTRS2015

Improving the Timeliness of Winter Wheat Production Forecast in the United States of America, Ukraine and China Using MODIS Data and NCAR Growing Degree Day Information

Wheat is themost important cereal crop traded on international markets and winter wheat constitutes approximately 80% of global wheat production. Thus, accurate and timely production forecasts are critical for making informed agricultural policies and investments, as well as increasing market efficiency and stability. Becker-Reshef et al. (2010) developed an empirical generalized model for forecasting winter wheat production. Their approach combined BRDF-corrected daily surface reflectance from Moderate resolution Imaging Spectroradiometer (MODIS) Climate Modeling Grid (CMG) with detailed official crop statistics and crop typemasks. It is based on the relationship between the Normalized Difference Vegetation Index (NDVI) at the peak of the growing season, percent wheat within the CMG pixel (area within the CMG pixel occupied by wheat), and the final yields. This method predicts the yield approximately one month to six weeks prior to harvest. In this study, we include Growing Degree Day (GDD) information extracted from NCEP/NCAR reanalysis data in order to improve the winter wheat production forecast by increasing the timeliness of the forecasts while conserving the accuracy of the original model. We apply this modified model to three major wheat-producing countries: the Unites States (US), Ukraine and China from 2001 to 2012. We show that a reliable forecast can be made between one month to a month and a half prior to the peak NDVI (meaning two months to two and a half months prior to harvest), while conserving an accuracy of 10% in the production forecast.

Production↗