Search NASA⌕ Search

SEARCH · Search NASA

Results for “mean squared error”

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 415 records · Page 23

Evaluation of interpolation methods for surface-based motion compensated tomographic reconstruction for cardiac angiographic C-arm data

Purpose: For interventional cardiac procedures, anatomical and functional information about the cardiac chambers is of major interest. With the technology of angiographic C-arm systems it is possible to reconstruct intraprocedural three-dimensional (3D) images from 2D rotational angiographic projection data (C-arm CT). However, 3D reconstruction of a dynamic object is a fundamental problem in C-arm CT reconstruction. The 2D projections are acquired over a scan time of several seconds, thus the projection data show different states of the heart. A standard FDK reconstruction algorithm would use all acquired data for a filtered backprojection and result in a motion-blurred image. In this approach, a motion compensated reconstruction algorithm requiring knowledge of the 3D heart motion is used. The motion is estimated from a previously presented 3D dynamic surface model. This dynamic surface model results in a sparse motion vector field (MVF) defined at control points. In order to perform a motion compensated reconstruction, a dense motion vector field is required. The dense MVF is generated by interpolation of the sparse MVF. Therefore, the influence of different motion interpolation methods on the reconstructed image quality is evaluated. Methods: Four different interpolation methods, thin-plate splines (TPS), Shepard's method, a smoothed weighting function, and a simple averaging, were evaluated. The reconstruction quality was measured on phantom data, a porcine model as well as on in vivo clinical data sets. As a quality index, the 2D overlap of the forward projected motion compensated reconstructed ventricle and the segmented 2D ventricle blood pool was quantitatively measured with the Dice similarity coefficient and the mean deviation between extracted ventricle contours. For the phantom data set, the normalized root mean square error (nRMSE) and the universal quality index (UQI) were also evaluated in 3D image space. Results: The quantitative evaluation of all experiments showed that TPS interpolation provided the best results. The quantitative results in the phantom experiments showed comparable nRMSE of Almost-Equal-To 0.047 {+-} 0.004 for the TPS and Shepard's method. Only slightly inferior results for the smoothed weighting function and the linear approach were achieved. The UQI resulted in a value of Almost-Equal-To 99% for all four interpolation methods. On clinical human data sets, the best results were clearly obtained with the TPS interpolation. The mean contour deviation between the TPS reconstruction and the standard FDK reconstruction improved in the three human cases by 1.52, 1.34, and 1.55 mm. The Dice coefficient showed less sensitivity with respect to variations in the ventricle boundary. Conclusions: In this work, the influence of different motion interpolation methods on left ventricle motion compensated tomographic reconstructions was investigated. The best quantitative reconstruction results of a phantom, a porcine, and human clinical data sets were achieved with the TPS approach. In general, the framework of motion estimation using a surface model and motion interpolation to a dense MVF provides the ability for tomographic reconstruction using a motion compensation technique.

60 APPLIED LIFE SCIENCES↗

Detection of signals by weighted integrate-and-dump filter

A Weighted Integrate and Dump Filter (WIDF) is presented that results in reducing those losses in telemetry symbol signal to noise ratio (SNR) which occur in digital Integrate and Dump Filters (IDFs) when the samples are not phase locked to the input data symbol clock. The Minimum Mean Square Error (MMSE) criterion is used to derive a set of weights for approximating the analog integrate and dump filter, which is the matched filter for detection of signals in additive white Gaussian noise. This new digital matched filter results in considerable performance improvement compared to unweighted digital matched filters. An example is presented for a sampling rate of four times the symbol rate. As the sampling offset (or phase) varies with respect to the data symbol boundaries, the output SNR varies 1 dB for an unweighted IDF, but only 0.3 dB for the optimum WIDF, averaged over random data patterns. This improvement in performance relative to unweighted IDF means that significantly lower sampling and processing rates can be used for given telemetry symbol rates, resulting in reduced system cost.

Sadr, R.↗

Technical Note: On the calculation of stopping-power ratio for stoichiometric calibration in proton therapy

Purpose: The quantitative effects of assumptions made in the calculation of stopping-power ratios (SPRs) are investigated, for stoichiometric CT calibration in proton therapy. The assumptions investigated include the use of the Bethe formula without correction terms, Bragg additivity, the choice of I-value for water, and the data source for elemental I-values. Methods: The predictions of the Bethe formula for SPR (no correction terms) were validated against more sophisticated calculations using the SRIM software package for 72 human tissues. A stoichiometric calibration was then performed at our hospital. SPR was calculated for the human tissues using either the assumption of simple Bragg additivity or the Seltzer-Berger rule (as used in ICRU Reports 37 and 49). In each case, the calculation was performed twice: First, by assuming the I-value of water was an experimentally based value of 78 eV (value proposed in Errata and Addenda for ICRU Report 73) and second, by recalculating the I-value theoretically. The discrepancy between predictions using ICRU elemental I-values and the commonly used tables of Janni was also investigated. Results: Errors due to neglecting the correction terms to the Bethe formula were calculated at less than 0.1% for biological tissues. Discrepancies greater than 1%, however, were estimated due to departures from simple Bragg additivity when a fixed I-value for water was imposed. When the I-value for water was calculated in a consistent manner to that for tissue, this disagreement was substantially reduced. The difference between SPR predictions when using Janni’s or ICRU tables for I-values was up to 1.6%. Experimental data used for materials of relevance to proton therapy suggest that the ICRU-derived values provide somewhat more accurate results (root-mean-square-error: 0.8% versus 1.6%). Conclusions: The conclusions from this study are that (1) the Bethe formula can be safely used for SPR calculations without correction terms; (2) simple Bragg additivity can be reasonably assumed for compound materials; (3) if simple Bragg additivity is assumed, then the I-value for water should be calculated in a consistent manner to that of the tissue of interest (rather than using an experimentally derived value); (4) the ICRU Report 37 I-values may provide a better agreement with experiment than Janni’s tables.

60 APPLIED LIFE SCIENCES↗

Resampling and data augmentation for short-term PV output prediction based on an imbalanced sky images dataset using convolutional neural networks

Integrating photovoltaics (PV) into electricity grids is challenged by potentially large fluctuations in power generation. In recent years, sky image-based PV output prediction using convolutional neural networks (CNNs) has emerged as a promising approach to forecasting fluctuations. A key challenge is imbalanced sky image datasets: because of the geography of solar PV system installations, sky image datasets are often rich in sunny condition data but deficient in cloudy condition data. This imbalance contrasts with the fact that model errors are dominated by cloudy condition performance. In this study, we attempt to remedy this by exploring the enrichment and augmentation of an imbalanced sky images dataset for two PV output prediction tasks: nowcasting (predicting concurrent PV output) and forecasting (predicting 15-minute-ahead future PV output). We empirically examine the efficacy of using different resampling and data augmentation approaches to create a rebalanced dataset for model development. A three-stage greedy search is used to determine the optimal resampling approach, data augmentation techniques and over-sampling rate. The results show that for the nowcast problem, resampling and data augmentation can effectively enhance the model performance, reducing overall root mean squared error (RMSE) by an average of 4%, or a 15 std. (standard deviation) of improvement compared to the variability of the baseline model. In contrast, the treatment RMSE for the forecast problem nearly always overlaps the baseline performance at the ± 2 std. level. The optimal resampling approach expands on the original dataset by over-sampling the minority cloudy data, with the best results from large over-sampling rate (e.g., 4 ~ 6 times over-sampling of cloudy images).

14 SOLAR ENERGY↗

The Effects of Aviation Weather Information Systems on General Aviation Weather Information Systems on General Pilots' Workload

My work at NASA Langley has focused around Aviation Weather Information CAWING displays. The majority of my time at LYRIC has been spent on the Workload and Relative Position (WaRP) Study. The goal of this project is to determine how an AWIN display at various positions within the cockpit affects pilot performance and workload. The project is being conducted in Languages Cessna 206H research aircraft. During the past year the design of the experiment was finalized and approved. Despite facing several delays the data collection was completed in early February. Alter the completion of the data collection an extensive data entry task began. This required recording air speed, altitude, course heading, bank angle, and vertical speed information from videos of the primary flight displays. This data was then used to determine root mean square error (RMSE) for each experimental condition. In addition to the performance data (RMSE) taken from flight path deviation, the study also collected data on pilot;s accuracy in reporting weather information, and a subjective rating of workload from the pilot. The data for this experiment is currently being analyzed. Overall the current experiment should help to determine potential costs and benefits associated with AWIN displays. The data will be used to determine if a private pilot can safely fly a general aviation aircraft while operating a weather display. Clearly a display that adds to the pilot#s already heavy workload represents a potential problem. The study will compare the use of an AWIN display to conventional means of acquiring weather data. The placement of the display within the cockpit (i.e., either on the yoke, kneeboard, or panel) will be also compared in terms of workload, performance, and pilot preference.

Scerbo, Mark↗

LSTM-Based Data Integration to Improve Snow Water Equivalent Prediction and Diagnose Error Sources

Accurate prediction of snow water equivalent (SWE) can be valuable for water resource managers. Recently, deep learning methods such as long short-term memory (LSTM) have exhibited high accuracy in simulating hydrologic variables and can integrate lagged observations to improve prediction, but their benefits were not clear for SWE simulations. Here we tested an LSTM network with data integration (DI) for SWE in the western United States to integrate 30-day-lagged or 7-day-lagged observations of either SWE or satellite-observed snow cover fraction (SCF) to improve future predictions. SCF proved beneficial only for shallow-snow sites during snowmelt, while lagged SWE integration significantly improved prediction accuracy for both shallow- and deep-snow sites. The median Nash–Sutcliffe model efficiency coefficient (NSE) in temporal testing improved from 0.92 to 0.97 with 30-day-lagged SWE integration, and root-mean-square error (RMSE) and the difference between estimated and observed peak SWE values d max were reduced by 41% and 57%, respectively. DI effectively mitigated accumulated model and forcing errors that would otherwise be persistent. Moreover, by applying DI to different observations (30-day-lagged, 7-day-lagged), we revealed the spatial distribution of errors with different persistent lengths. For example, integrating 30-day-lagged SWE was ineffective for ephemeral snow sites in the southwestern United States, but significantly reduced monthly-scale biases for regions with stable seasonal snowpack such as high-elevation sites in California. These biases are likely attributable to large interannual variability in snowfall or site-specific snow redistribution patterns that can accumulate to impactful levels over time for nonephemeral sites. These results set up benchmark levels and provide guidance for future model improvement strategies.

54 ENVIRONMENTAL SCIENCES↗

Emulator-Based Bayesian Calibration of the CISNET Colorectal Cancer Models

Purpose To calibrate Cancer Intervention and Surveillance Modeling Network (CISNET)'s SimCRC, MISCAN-Colon, and CRC-SPIN simulation models of the natural history colorectal cancer (CRC) with an emulator-based Bayesian algorithm and internally validate the model-predicted outcomes to calibration targets.Methods We used Latin hypercube sampling to sample up to 50,000 parameter sets for each CISNET-CRC model and generated the corresponding outputs. We trained multilayer perceptron artificial neural networks (ANNs) as emulators using the input and output samples for each CISNET-CRC model. We selected ANN structures with corresponding hyperparameters (i.e., number of hidden layers, nodes, activation functions, epochs, and optimizer) that minimize the predicted mean square error on the validation sample. We implemented the ANN emulators in a probabilistic programming language and calibrated the input parameters with Hamiltonian Monte Carlo-based algorithms to obtain the joint posterior distributions of the CISNET-CRC models' parameters. We internally validated each calibrated emulator by comparing the model-predicted posterior outputs against the calibration targets.Results The optimal ANN for SimCRC had 4 hidden layers and 360 hidden nodes, MISCAN-Colon had 4 hidden layers and 114 hidden nodes, and CRC-SPIN had 1 hidden layer and 140 hidden nodes. The total time for training and calibrating the emulators was 7.3, 4.0, and 0.66 h for SimCRC, MISCAN-Colon, and CRC-SPIN, respectively. The mean of the model-predicted outputs fell within the 95% confidence intervals of the calibration targets in 98 of 110 for SimCRC, 65 of 93 for MISCAN, and 31 of 41 targets for CRC-SPIN.Conclusions Using ANN emulators is a practical solution to reduce the computational burden and complexity for Bayesian calibration of individual-level simulation models used for policy analysis, such as the CISNET CRC models. In this work, we present a step-by-step guide to constructing emulators for calibrating 3 realistic CRC individual-level models using a Bayesian approach.

artificial neural networks↗

US East Coast synthetic aperture radar wind atlas for offshore wind energy

We present the first synthetic aperture radar (SAR) offshore wind atlas of the US East Coast from Georgia to the Canadian border. Images from RADARSAT-1, Envisat, and Sentinel-1A/B are processed to wind maps using the geophysical model function (GMF) CMOD5.N. Extensive comparisons with 6008 collocated buoy observations of the wind speed reveal that biases of the individual systems range from -0.8 to 0.6 m s -1 . Unbiased wind retrievals are crucial for producing an accurate wind atlas, and intercalibration of the SAR observations is therefore applied. Wind retrievals from the intercalibrated SAR observations show biases in the range of to -0.2 to 0.0 m s -1 , while at the same time improving the root-mean-squared error from 1.67 to 1.46 m s -1 . The intercalibrated SAR observations are, for the first time, aggregated to create a wind atlas at the height 10 m a.s.l. (above sea level). The SAR wind atlas is used as a reference to study wind resources derived from the Wind Integration National Dataset Toolkit (WTK), which is based on 7 years of modelling output from the Weather Research and Forecasting (WRF) model. Comparisons focus on the spatial variation in wind resources and show that model outputs lead to lower coastal wind speed gradients than those derived from SAR. Areas designated for offshore wind development by the Bureau of Ocean Energy Management are investigated in more detail; the wind resources in terms of the mean wind speed show spatial variations within each designated area between 0.3 and 0.5 m s -1 for SAR and less than 0.2m s -1 for the WTK. Our findings indicate that wind speed gradients and variations might be underestimated in mesoscale model outputs along the US East Coast.

17 WIND ENERGY↗

Performance of methods for SARS-CoV-2 variant detection and abundance estimation within mixed population samples

The accurate identification of SARS-CoV-2 (SC2) variants and estimation of their abundance in mixed population samples (e.g., air or wastewater) is imperative for successful surveillance of community level trends. Assessing the performance of SC2 variant composition estimators (VCEs) should improve our confidence in public health decision making. Here, we introduce a linear regression based VCE and compare its performance to four other VCEs: two re-purposed DNA sequence read classifiers (Kallisto and Kraken2), a maximum-likelihood based method (Lineage deComposition for Sars-Cov-2 pooled samples (LCS)), and a regression based method (Freyja). We simulated DNA sequence datasets of known variant composition from both Illumina and Oxford Nanopore Technologies (ONT) platforms and assessed the performance of each VCE. We also evaluated VCEs performance using publicly available empirical wastewater samples collected for SC2 surveillance efforts. Bioinformatic analyses were performed with a custom NextFlow workflow (C-WAP, CFSAN Wastewater Analysis Pipeline). Relative root mean squared error (RRMSE) was used as a measure of performance with respect to the known abundance and concordance correlation coefficient (CCC) was used to measure agreement between pairs of estimators. Based on our results from simulated data, Kallisto was the most accurate estimator as it had the lowest RRMSE, followed by Freyja. Kallisto and Freyja had the most similar predictions, reflected by the highest CCC metrics. We also found that accuracy was platform and amplicon panel dependent. For example, the accuracy of Freyja was significantly higher with Illumina data compared to ONT data; performance of Kallisto was best with ARTICv4. However, when analyzing empirical data there was poor agreement among methods and variations in the number of variants detected (e.g., Freyja ARTICv4 had a mean of 2.2 variants while Kallisto ARTICv4 had a mean of 10.1 variants). This work provides an understanding of the differences in performance of a number of VCEs and how accurate they are in capturing the relative abundance of SC2 variants within a mixed sample (e.g., wastewater). Such information should help officials gauge the confidence they can have in such data for informing public health decisions.

60 APPLIED LIFE SCIENCES↗

Regional Characteristics of the Relationship Between Columnar AOD and Surface PM2.5: Application of Lidar Aerosol Extinction Profiles over Baltimore-Washington Corridor During DISCOVER-AQ

The first field campaign of DISCOVER-AQ (Deriving Information on Surface conditions from COlumn and VERtically resolved observations relevant to Air Quality) took place in July 2011 over Baltimore-Washington Corridor (BWC). A suite of airborne remote sensing and in-situ sensors was deployed along with ground networks for mapping vertical and horizontal distribution of aerosols. Previous researches were based on a single lidar station because of the lack of regional coverage. This study uses the unique airborne HSRL (High Spectral Resolution Lidar) data to baseline PM2.5 (particulate matter of aerodynamic diameter less than 2.5 μm) estimates and applies to regional air quality with satellite AOD (Aerosol Optical Depth) retrievals over BWC (∼6500 sq. km). The linear approximation takes into account aerosols aloft above AML (Aerosol Mixing Layer) by normalizing AOD with haze layer height (i.e., AOD/HLH). The estimated PM2.5 mass concentrations by HSRL AOD/HLH are shown within 2 RMSE (Root Mean Square Error ∼9.6 μg/cu. m) with correlation ∼0.88 with the observed over BWC. Similar statistics are shown when applying HLH data from a single location over the distance of 100 km. In other words, a single lidar is feasible to cover the range of 100 km with expected uncertainties. The employment of MPLNET-AERONET (MicroPulse Lidar NETwork - AErosol RObotic NETwork) measurements at NASA GSFC produces similar statistics of PM2.5 estimates as those derived by HSRL. The synergy of active and passive remote sensing aerosol measurements provides the foundation for satellite application of air quality on a daily basis. For the optimal range of 10 km, the MODIS-estimated PM2.5 values are found satisfactory at 27 (out of 36) sunphotometer locations with mean RMSE of 1.6-3.3 μg/cu. m relative to PM2.5 estimated by sunphotometers. The remaining 6 of 8 marginal sites are found in the coastal zone, for which associated large RMSE values ∼4.5-7.8 μg/cu. m are most likely due to overestimated AOD because of water-contaminated pixels.

DISCOVER-AQ↗

Development of an open-source regional data assimilation system in PEcAn v. 1.7.2: application to carbon cycle reanalysis across the contiguous US using SIPNET

Abstract. The ability to monitor, understand, and predict the dynamics of the terrestrial carbon cycle requires the capacity to robustly and coherently synthesize multiple streams of information that each provide partial information about different pools and fluxes. In this study, we introduce a new terrestrial carbon cycle data assimilation system, built on the PEcAn model–data eco-informatics system, and its application for the development of a proof-of-concept carbon “reanalysis” product that harmonizes carbon pools (leaf, wood, soil) and fluxes (GPP, Ra, Rh, NEE) across the contiguous United States from 1986–2019. We first calibrated this system against plant trait and flux tower net ecosystem exchange (NEE) using a novel emulated hierarchical Bayesian approach. Next, we extended the Tobit–Wishart ensemble filter (TWEnF) state data assimilation (SDA) framework, a generalization of the common ensemble Kalman filter which accounts for censored data and provides a fully Bayesian estimate of model process error, to a regional-scale system with a calibrated localization. Combined with additional workflows for propagating parameter, initial condition, and driver uncertainty, this represents the most complete and robust uncertainty accounting available for terrestrial carbon models. Our initial reanalysis was run on an irregular grid of ∼ 500 points selected using a stratified sampling method to efficiently capture environmental heterogeneity. Remotely sensed observations of aboveground biomass (Landsat LandTrendr) and leaf area index (LAI) (MODIS MOD15) were sequentially assimilated into the SIPNET model. Reanalysis soil carbon, which was indirectly constrained based on modeled covariances, showed general agreement with SoilGrids, an independent soil carbon data product. Reanalysis NEE, which was constrained based on posterior ensemble weights, also showed good agreement with eddy flux tower NEE and reduced root mean square error (RMSE) compared to the calibrated forecast. Ultimately, PEcAn's new open-source regional data assimilation framework provides a scalable workflow for harmonizing multiple data constraints and providing a uniform synthetic platform for carbon monitoring, reporting, and verification (MRV) as well as accelerating terrestrial carbon cycle research.

54 ENVIRONMENTAL SCIENCES↗

HAPPA: A Modular Platform for HPC Application Resilience Analysis with LLMs Embedded

High-performance computing (HPC) systems are increasingly vulnerable to soft errors, which pose significant challenges in maintaining computational accuracy and reliability. Predicting the resilience of HPC applications to these errors is crucial for robust code protection and detailed resilience analysis. In this study, we present HAppA, a modular platform designed for HPC Application Resilience Analysis. Embedding Large Language Models (LLMs), HAppA addresses understanding the context information of long code sequences typical in HPC applications. HAppA implements a novel code representation module that chunks the code into fixed-size segments and aggregates the embeddings of these segments. Three aggregation methods have been explored: MeanPooling, MaxPooling, and LSTM-based techniques. We built a DAtaset for REsilience analysis using Fault Injection (FI), named DARE. Using our DARE dataset, HAppA is trained for regression prediction tasks. Our evaluation results demonstrate the predictive accuracy of HAppA compared to other models, particularly noting that the LSTM-based aggregation method -- HAppA-LSTM -- achieves a mean squared error (MSE) of 0.078 for SDC prediction, surpassing the existing state-of-the-art PARIS model, which recorded an MSE of 0.1172. Additionally, HAppA with the KeyBERT model extracts a list of keywords representing the source code. A comprehensive importance analysis of these keywords further elucidates the code patterns contributing to the error rate. These findings highlight the effectiveness of HAppA in analyzing the resilience of HPC applications and establish a new benchmark for predictive accuracy in resilience.

Jiang, Hailong [Kent State University]↗

An L-Band Brightness Temperature Disaggregation Method Using S-Band Radiometer Data for the Water Cycle Observation Mission (WCOM)

The Water Cycle Observation Mission (WCOM) will build upon previous L and C band passive microwave soil moisture satellite missions. WCOM will consist of a passive microwave synthetic aperture radiometer operating at L, S, and C bands. The WCOM requirements for passive soil moisture are to estimate soil moisture in the top 5 cm of soil layer with an error less than 0.04 cu m/cu m, at 15 km resolution and with a 3-day revisit. A new set of algorithms for these multi-frequency platforms will need to be developed for estimating the data products at the desired resolution. To accomplish this, a brightness temperature (TB) downscaling methodology is developed that uses passive S-band TB (30 km) to downscale L-band TB (50 km) and to estimate soil moisture at a 30 km resolution, based on the linear relationships between the passive signals of L-band and S-band. To test this downscaling method, analysis was performed using PALS data from the Soil Moisture Experiments in 2002 (SMEX02). For this study, 4 km L-band observations were downscaled to 800 m. The root mean square errors (RMSE) between the downscaled TBL at 800m with the observed TBL at 800m are 2.63 K and 1.60 K for H and V polarizations respectively. The results also showed that it was possible to use these disaggregated TB to estimate soil moisture to meet the mission requirement of 0.04 cu m/cu m. These results showed that we can obtain higher resolution soil moisture from L band passive TB with a high accuracy (<0.04 cu m/cu m) by using S-band information.

Disaggregation↗

WE-G-207-02: Full Sequential Projection Onto Convex Sets (FS-POCS) for X-Ray CT Reconstruction

Purpose: To develop an iterative reconstruction method for X-ray CT, in which the reconstruction can quickly converge to the desired solution with much reduced projection views. Methods: The reconstruction is formulated as a convex feasibility problem, i.e. the solution is an intersection of three convex sets: 1) data fidelity (DF) set – the L2 norm of the difference of observed projections and those from the reconstructed image is no greater than an error bound; 2) non-negativity of image voxels (NN) set; and 3) piecewise constant (PC) set - the total variation (TV) of the reconstructed image is no greater than an upper bound. The solution can be found by applying projection onto convex sets (POCS) sequentially for these three convex sets. Specifically, the algebraic reconstruction technique and setting negative voxels as zero are used for projection onto the DF and NN sets, respectively, while the projection onto the PC set is achieved by solving a standard Rudin, Osher, and Fatemi (ROF) model. The proposed method is named as full sequential POCS (FS-POCS), which is tested using the Shepp-Logan phantom and the Catphan600 phantom and compared with two similar algorithms, TV-POCS and CP-TV. Results: Using the Shepp-Logan phantom, the root mean square error (RMSE) of reconstructed images changing along with the number of iterations is used as the convergence measurement. In general, FS- POCS converges faster than TV-POCS and CP-TV, especially with fewer projection views. FS-POCS can also achieve accurate reconstruction of cone-beam CT of the Catphan600 phantom using only 54 views, comparable to that of FDK using 364 views. Conclusion: We developed an efficient iterative reconstruction for sparse-view CT using full sequential POCS. The simulation and physical phantom data demonstrated the computational efficiency and effectiveness of FS-POCS.

60 APPLIED LIFE SCIENCES↗

Use of In Situ and Airborne Multiangle Data to Assess MODIS and Landsat-Based Estimates of Directional Reflectance and Albedo

The quantification of uncertainty in satellite-derived global surface albedo products is a critical aspect in producing complete, physically consistent, and decadal land property data records for studying ecosystem change. A challenge in validating albedo measurements acquired from space is the ability to overcome the spatial scaling errors that can produce disagreements between satellite and field-measured values. Here, we present the results from an accuracy assessment of MODIS and Landsat-TM albedo retrievals, based on collocated comparisons with tower and airborne Cloud Absorption Radiometer (CAR) measurements collected during the 2007 Cloud and Land Surface Interaction Campaign (CLASIC). The initial focus was on evaluating intersensor consistency through comparisons of intrinsic bidirectional reflectance estimates. Local and regional assessments were then performed to obtain estimates of the resulting scaling uncertainties, and to establish the accuracy of albedo reconstructions during extended periods of precipitation. In general, the satellite derived estimates met the accuracy requirements established for the high-quality MODIS operational albedos at 500 m (the greater of 0.02 units or +/-10% of surface measured values). However, results reveal a high degree of variability in the root-mean-square error (RMSE) and bias of MODIS visible (0.3-0.7 μm) and Landsat-TM shortwave (0.3-5.0 μm) albedos; where, in some cases, retrieval uncertainties were found to be in excess of 15%. Results suggest that an overall improvement in MODIS shortwave albedo retrieval accuracy of 7.8%, based on comparisons between MODIS and CAR albedos, resulted from the removal of sub-grid scale mismatch errors when directly scaling-up the tower measurements to the MODIS satellite footprint.

biosphere↗

Estimation of Corn Latent Heat Flux from High Resolution Thermal Imagery

Crop evapotranspiration (ET), which is directly related to latent heat flux, is also a key indicator in determining the water status of crops. In order to estimate the latent heat flux, two-source energy balance (TSEB) models have been developed for thermal imagery from satellite platforms. However, because of the coarse resolution of thermal sensors on the satellite, distinguishing soil and vegetation is difficult which complicates the calculation process and introduces errors in latent heat estimates. In this research, high-resolution thermal datasets (0.05 m) and corresponding RGB datasets (0.03 m) were used for calculating crop latent heat flux using an adapted TSEB model. The RGB datasets were used for supervised classification of soil and vegetation, and the classification results were then used to filter the thermal mosaics to separate vegetation and soil temperatures. The vegetation temperature is used for calculating latent heat flux and the results are validated against the ground reference measurements of latent heat using a handheld porometer. The objective of this research is to introduce a workflow including an adapted TSEB model which is customized for high resolution thermal images from unmanned aircraft systems (UAS) to estimate the latent heat flux of row crops in agricultural fields. Nine dates of data collection in 2018 and 2020 have been evaluated and the root mean square error (RMSE) varies between 16 to 106 W/m2 depending on the days after planting (DAP) and the time of measurement for each day. The results indicate that the workflow introduced here is able to provide estimates of instantaneous latent heat flux (evapotranspiration) measurements for row crops in agricultural fields which will enable people to make reliable decisions related to irrigation scheduling.

54 ENVIRONMENTAL SCIENCES↗

A Coupled Deep Learning Model for Estimating Surface NO 2 Levels from Remote Sensing Data: 15-Year Study Over the Contiguous United States

This study proposes a novel two-step deep learning (DL) model for estimating surface NO 2 concentrations using satellite data over the contiguous United States (CONUS) from 2005 to 2019. The first phase of the model uses partial convolutional neural network (PCNN), an advanced DL model that accurately imputes gaps between surface NO 2 stations and creates 5,478 daily-mean NO 2 grids (PCNN-NO 2 ) of the 2005-2019 period over the study area. We then feed the PCNN-NO 2 , along with other predictor variables, into a deep neural network (DNN) to estimate surface NO 2 levels, achieving exceptional performance with a correlation coefficient of 0.975 to 0.978, a mean absolute bias of 0.99 ppb to 1.38 ppb, and a root mean square error of 1.47 ppb to 1.97 ppb. Spatial cross-validation results also indicate strong spatial performance of PCNN-DNN surface NO 2 estimates. In addition to its accurate estimates, the PCNN-DNN model consistently generates estimated NO 2 grids without any missing values, improving the quality of various applications such as emission reduction strategies and public health studies. Between 2005 and 2019, the 5,478 daily estimated NO 2 grids over the CONUS reveal significant reductions in NO 2 levels in fourteen major urban environments: Washington D.C. (-43%), New York (-45%), Los Angeles (-38%), Chicago (-25%), Boston (-43%), Houston (-34%), Dallas (-40%), Philadelphia (-41%), Phoenix (-38%), Detroit (-20%), Denver (-23%), Atlanta (-0.7%), Cincinnati (-38%), and Pittsburgh (-56%). Furthermore, the study shows that the denser urban regions that in-situ stations are installed in, the higher the difference between in-situ observations and regional-mean NO 2 levels.

54 ENVIRONMENTAL SCIENCES↗

Assessing the Viability of Using GEOS-Forecast Product for Landslides Forecasting: A Step Toward Early Warning System

Landslides across the globe are mostly triggered by extreme rainfall events affecting infrastructure, transportation and livelihoods. The risks are rarely quantified due to lack of data, analytical skills and limited modeling techniques. Knowledge of local to global scale landslide risks provides communities and national agencies the ability to adapt disaster management practices to mitigate and recover from these hazards. In order to minimize the risks and improve characterization of community resilience to landslides, it is vital to have reliable information about the factors triggering landslides such as rainfall, well ahead in time. Forecasting potential landslide activity and impacts can be achieved through reliable precipitation forecast models. However, it is challenging because of the temporal and spatial variability of precipitation, an important factor in triggering landslides. Evaluation of the precipitation field, associated errors, and sampling uncertainties is integral for development of efficient and reliable landslide forecasting and early warning system. This study develops a methodology to assess the viability of using a precipitation field provided by a global model and its potential integration in the landslide forecasting system. The study focuses on the comparison between the IMERG (Integrated Multi-satellitE Retrievals for Global Precipitation Mission) and GEOS (NASA Goddard Earth Observing System)-Forecast product over contiguous United States (CONUS). GEOS model assimilates new observations every 6 hours, at 00, 06, 12, and 18 UTC. The framework is tested on the GEOS-Forecast Model initialized at 00 UTC using daily IMERG early product as reference using both categorical and continuous statistics. The categorical statistics includes the probability of detection (POD), success ratio (SR), critical success index (CSI), and the hit bias. Continuous statistics such as correlation, normalized standard deviation, and root-mean-square error are also evaluated. Overall, GEOS-Forecast precipitation field over the analysis period (~1 year) show underestimation with respect to IMERG early for the daily accumulated rainfall. However, the probability distribution function and cumulative distribution function of both show similar patterns. In terms of correlations, POD, SR, CSI, hit bias, the performance varies with respect to the rainfall threshold used.

Sana Khan↗