SEARCH · Engineering Papers
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.
Statistical prediction methods for short-term (months) and long-term (years) forecasting of solar activity were studied. The comparisons indicate that better predictions, in a chi square sense, are possible by lining up the maximum (or minimums, or both) by cycle number. Evidence is also presented to support the existence of an aperiodic variation in the periods as well as the amplitudes.
Two major themes were pursued during this research period. The first of these involved examining the impacts of satellite-based data and the forecast model used by the Goddard Laboratory for Atmospheres (GLA) on general circulation statistics. For the other major topic, the diabatic heating fields produced by GLA were examined for one month during the FGGE First Special Observing Period. As part of that effort, the three-dimensional distribution of the four component heating fields were studied, namely those due to shortwave radiation, Q sub SW, longwave radiation, Q sub LW, sensible heating, Q sub S, and latent heating, Q sub L. These components were calculated as part of the GLA analysis/forecast system and archived every quarter day; from these archives cross products with temperature were computed to enable the direct calculation of certain terms of the large-scale atmospheric energy cycle, namely those involving the generation of available potential energy (APE). The decision to archive the diabatic heating components separately has enabled researchers to study the role of the various processes that drive the energy cycle of the atmosphere.
The computational complexity of algorithms for Four Dimensional Data Assimilation (4DDA) at NASA's Data Assimilation Office (DAO) is discussed. In 4DDA, observations are assimilated with the output of a dynamical model to generate best-estimates of the states of the system. It is thus a mapping problem, whereby scattered observations are converted into regular accurate maps of wind, temperature, moisture and other variables. The DAO is developing and using 4DDA algorithms that provide these datasets, or analyses, in support of Earth System Science research. Two large-scale algorithms are discussed. The first approach, the Goddard Earth Observing System Data Assimilation System (GEOS DAS), uses an atmospheric general circulation model (GCM) and an observation-space based analysis system, the Physical-space Statistical Analysis System (PSAS). GEOS DAS is very similar to global meteorological weather forecasting data assimilation systems, but is used at NASA for climate research. Systems of this size typically run at between 1 and 20 gigaflop/s. The second approach, the Kalman filter, uses a more consistent algorithm to determine the forecast error covariance matrix than does GEOS DAS. For atmospheric assimilation, the gridded dynamical fields typically have More than 10(exp 6) variables, therefore the full error covariance matrix may be in excess of a teraword. For the Kalman filter this problem can easily scale to petaflop/s proportions. We discuss the computational complexity of GEOS DAS and our implementation of the Kalman filter. We also discuss and quantify some of the technical issues and limitations in developing efficient, in terms of wall clock time, and scalable parallel implementations of the algorithms.
The Integrated Medical Model (IMM) captures organizational knowledge across the space medicine, training, operations, engineering, and research domains. IMM uses this knowledge in the context of a mission and crew profile to forecast risks to crew health and mission success. The IMM establishes a quantified, statistical relationship among medical conditions, risk factors, available medical resources, and crew health and mission outcomes. These relationships may provide an appropriate foundation for developing an in-flight medical decision support tool that helps optimize the use of medical resources and assists in overall crew health management by an autonomous crew with extremely limited interactions with ground support personnel and no chance of resupply.
Grid-connected power electronic systems require timely and reliable fault detection to prevent equipment damage and reduce downtime. This paper presents a forecasting-based anomaly detection pipeline that decomposes voltage and current measurements into intrinsic mode functions (IMFs) using empirical mode decomposition (EMD), then trains a causal temporal convolutional network (TCN) on normal-operation IMF data to predict short-horizon future dynamics. Deviations between forecasts and observations are summarized as reliability-weighted residual scores and thresholded per sensor using robust statistics with temporal persistence constraints to suppress false positives. To reduce runtime, EMD is performed on downsampled signals for detection, while raw-rate EMD is applied only within a short region of interest for high-frequency interpretability near detected events. Results on a simulated grid-connected converter system demonstrate that IMF-domain forecasting improves anomaly separability relative to raw-signal forecasting and provides interpretable evidence of faults across decomposition channels.
Models for forecasting rainfall-induced landslides are mostly based on the identification of empirical rainfall thresholds obtained exploiting rain gauge data. Despite their increased availability, satellite rainfall estimates are scarcely used for this purpose. Satellite data should be useful in ungauged and remote areas, or should provide a significant spatial and temporal reference in gauged areas. In this paper, the analysis of the reliability of rainfall thresholds based on rainfall remote sensed and rain gauge data for the prediction of landslide occurrence is carried out. To date, the estimation of the uncertainty associated with the empirical rainfall thresholds is mostly based on a bootstrap resampling of the rainfall duration and the cumulated event rainfall pairs (D,E) characterizing rainfall events responsible for past failures. This estimation does not consider the measurement uncertainty associated with D and E. In the paper, we propose (i) a new automated procedure to reconstruct ED conditions responsible for the landslide triggering and their uncertainties, and (ii) three new methods to identify rainfall threshold for the possible landslide occurrence, exploiting rain gauge and satellite data. In particular, the proposed methods are based on Least Square (LS), Quantile Regression (QR) and Nonlinear Least Square (NLS) statistical approaches. We applied the new procedure and methods to define empirical rainfall thresholds and their associated uncertainties in the Umbria region (central Italy) using both rain-gauge measurements and satellite estimates. We finally validated the thresholds and tested the effectiveness of the different threshold definition methods with independent landslide information. The NLS method among the others performed better in calculating thresholds in the full range of rainfall durations. We found that the thresholds obtained from satellite data are lower than those obtained from rain gauge measurements. This is in agreement with the literature, where satellite rainfall data underestimate the 'ground' rainfall registered by rain gauges.
Argentina is one of the major producers and exporter of soybeans, corn, and wheat to the world market; therefore, the accurate and timely forecasting of those crops yield is crucial to national crop management and global food security. Previous studies have mainly focused on developing forecasting models for a specific crop type and location using a single source of data (e.g., vegetation indices), thus providing little insight into the forecasting models' performance on different crop types and regions. Besides, these models are based on traditional statistical regression algorithms, while more advanced machine learning approaches have not been explored. This study investigated the estimation of crop yields of three major crops (corn, soybean, and winter wheat) using Multiple Linear Regression (MLR) and Support Vector Machine (SVM), over major growing provinces in Argentina. Our models were trained and evaluated on data from 2015 to 2020, where three remote sensing products (Normalized difference vegetation index (NDVI), SMAP soil moisture, and MODIS evapotranspiration) were used as predictors. Our results indicated that accurate crop yield forecasts using the developed regression models could be made one to two months before harvest. The MLR and SVM model performance varied among different crop types, where soybean and corn exhibited better predictability compare to the wheat. In most cases, the SVM outperformed the multiple linear regression model due to its ability to capture the nonlinear and complex features of the crop-production process. The forecasted model that combines data from multiple sources outperformed single-source satellite data. The highest accuracy was obtained when the three data sources were all considered in the model development. Results also indicated that the inclusion of SMAP soil moisture improved crop yield forecasting in most provinces, and the most significant improvements occurred in the drier region.
Electricity load forecasting is essential to utility operation and power grid stability. A wide spectrum of data-driven methods, ranging from linear regression models to more recent deep learning models have been adopted to forecast electric load over the years. However, there still lacks a holistic evaluation of the applicability of conventional statistical and machine learning based algorithms with respect to different temporal and spatial scopes, computational requirements, and sensitivity of model-tuning. Enabled by a large-scale electricity load profile dataset of over 40,000 residential customers in a utility region, we conducted a cross-dimensional analysis of data-driven load forecasting methods. Three regression-based and seven deep learning algorithms with different model configurations were evaluated in terms of their overall and peak load prediction accuracy, and training burdens, across spatial aggregation levels ranging from the transformer, feeder, substation, to neighborhood. We found, first, the load forecasting accuracy is constrained by a predictability boundary, influenced by the forecasting horizon and spatial aggregation level. Specifically, RandomForest, XGBoost, TFT, TSMixer, and TiDE models achieved less than 10 % prediction error for up to 96-h ahead forecasting for district, substation, and feeder levels, while other models struggle at long-horizon predictions; Second, for winter and summer peak load dates, most models were able to predict the peak demand timing within ± 1 h, but the prediction percentage error varied by models, with TFT and TiDE models being the top performers; Third, models with similar prediction accuracy can differ in training burden by an order of magnitude. Therefore, choosing model configurations that balance prediction performance and computational resource is an important practical consideration for large-scale deployment of the machine learning based load forecasting. The outcome of this study can guide researchers and practitioners to choose the proper load forecasting algorithms based on their problem scope, required accuracy, and available resources. The predictability boundary can serve as a benchmark for electricity load forecasting problems with new algorithms and datasets.
In recent years, the Artificial Intelligence and Data Mining (AI&DM) models have become popular tools in assisting various aspects of reservoir operation. However, the practical uses are still rarely reported. Comparison experiment of many AI&DM models over a large number of reservoir cases is particularly valuable to help reservoir operators first examine the usefulness and transferability of different AI&DM models, and then identify the most stable and reliable AI&DM model in assist of various decision-making processes. In this study, a total of 12 AI&DM models with different parameterizations and simulation scenarios are comprehensively tested out and compared in simulating the controlled reservoir outflows of 33 reservoir cases over the Upper Colorado Region, United States. Results show that the Random Forecast and the Long-Short-Term-Memory model could consistently derive the best statistical performance than other models under the baseline simulation scenario. The employed AI&DM models could obtain satisfactory statistical interquartile ranges (25–75%) between [0.6–0.9], [0.3–0.8], and [0.2–0.8], for CORR, NSE, and KGE measurements, respectively, and [1.5–6.5], [–15 to 20], and [0.5–8.5] for the normalized RMSE, PBIAS and RSR measurements, respectively. Results also show Multi-Layer Perceptron model and Extreme Gradient Boosting Tree Algorithm produced more stable and superior performance than other models under more complex input scenarios. We also found that the performance of different AI&DM models are closely relevant to the reservoir elevations, sizes, and functionalities. Discussions were made about the sensitivity of AI&DM models’ parameterizations and the key advantages of AI&DM models over the rule-based reservoir models. We further identify that the main advantage of AI&DM models is the flexibility in designing input structures, whereas the rule-based simulation model is rather limited. Future studies were suggested regarding the best way reservoir operators and researchers could use, select, and apply different AI&DM models in simulating reservoir releases under different natural and modeling environments. Finally, this comparison study also serves as a reference and a piece of groundwork for further promoting the practical uses of AI&DM models in assisting reservoir operation.
The role of transient eddies in maintaining the observed local seasonal mean atmospheric circulation was investigated by examining the time-averaged momentum balances and omega equation, using seasonal statistics calculated from daily operational analyses by the European Centre for Medium Range Weather Forecasts. While both the Northern and Southern Hemispheres and several seasons were studied, emphasis was placed upon the Northern Hemisphere during December 1981-February 1982. The results showed that transient eddies played a secondary role in the seasonal mean zonal momentum budget and in the forcing of seasonal mean vertical and a geostrophic motion.
Seasonal statistics calculated from daily operational analyses by the European Centre for Medium Range Weather Forecasts for December 1980 - February 1981 (season 1) and December 1981 - February 1982 (season 2) were used to contrast the two seasonal mean circulation patterns present in the Northern Hemisphere and to investigate possible causes for the differences. The vertically averaged seasonal mean zonal wind and E-vector, a measure of the anisotropy of the transient eddies over the Pacific is shown for season 1 and season 2. The pattern in season 2, but not season 1, resembles that found in a modeling study of barotropic instability. One possible explanation for the differences between season 1 and 2 is a weak change in tropical heating associated with the southern oscillation, the effect of which in midlatitude was amplified by the occurrence of barotropic instability over the Pacific in season 2, but not season 1.
Seasonal statistics calculated from daily operational analyses by the European Center for Medium Range Weather Forecasts for December 1980 to February 1981 (season 1) and December 1981 to February 1982 (season 2) were used to contrast the two seasonal mean circulation patterns present in the Northern Hemisphere and to investigate possible causes for the differences. The vertically averaged seasonal mean zonal wind and E-vector, a measure of the anisotropy of the transient eddies over the Pacific is shown for season 1 and season 2. The pattern in season 2, but not season 1, resembles that found in a modeling study of barotropic instability. One possible explanation for the differences between season 1 and 2 is a weak change in tropical heating associated with the southern oscillation, the effect of which in midlatitude was amplified by the occurrence of barotropic instability over the Pacific in season 2, but not season 1.
The Arctic Region Supercomputing Center has been running a quasi-operational prototype of a High Resolution Rapid Refresh for Alaska (HRRRAK) at 3km resolution, initialized by the 13km Rapid Refresh (RR). Although the RR assimilates a broad range of observations into its analyses, experiments with the HRRRAK suggest that there may be added value in assimilating observations into the 3km initial conditions, downscaled from the 13km RR analyses. The NASA Short-term Prediction Research and Transition (SPoRT) group has been using assimilated data from the Atmospheric Infrared Sounder (AIRS) in WRF and WRF-Var simulations since 2004 with promising results. The sounder is aboard NASA s Aqua satellite, and provides vertical profiles of temperature and humidity. The Gridpoint Statistical Interpolation (GSI) system is then used to assimilate these vertical profiles into WRF forecasts. In this work, we assess the use of AIRS data in combination with other global data assimilation products on non-assimilated HRRRAK case studies. Two separate weather events will be assessed to qualitatively and quantitatively assess the impacts of AIRS data on HRRRAK forecasts.
The Arctic Region Supercomputing Center has been running a quasi-operational prototype of a High Resolution Rapid Refresh for Alaska (HRRRAK) at 3km resolution, initialized by the 13km Rapid Refresh (RR). Although the RR assimilates a broad range of observations into its analyses, experiments with the HRRRAK suggest that there may be added value in assimilating observations into the 3km initial conditions, downscaled from the 13km RR analyses. The NASA Short-term Prediction Research and Transition (SPoRT) group has been using assimilated data from the Atmospheric Infrared Sounder (AIRS) in WRF and WRF-Var simulations since 2004 with promising results. The sounder is aboard NASA s Aqua satellite, and provides vertical profiles of temperature and humidity. The Gridpoint Statistical Interpolation (GSI) system is then used to assimilate these vertical profiles into WRF forecasts. In this work, we assess the use of AIRS data in combination with other global data assimilation products on non-assimilated HRRRAK case studies. Two separate weather events will be assessed to qualitatively and quantitatively assess the impacts of AIRS data on HRRRAK forecasts.
This report summarizes the Applied Meteorology Unit (AMU) activities for the third quarter of Fiscal Year 2008 (April - June 2008). Tasks reported on are: Peak Wind Tool for User Launch Commit Criteria (LCC), Anvil Forecast Tool in AWIPS Phase II, Completion of the Edward Air Force Base (EAFB) Statistical Guidance Wind Tool, Volume Averaged Height Integ rated Radar Reflectivity (VAHIRR), Impact of Local Sensors, Radar Scan Strategies for the PAFB WSR-74C Replacement, VAHIRR Cost Benefit Analysis, and WRF Wind Sensitivity Study at Edwards Air Force Base
Predictions of solar activity are an essential part of our Space Weather forecast capability. Users are requiring usable predictions of an upcoming solar cycle to be delivered several years before solar minimum. A set of predictions of the amplitude of Solar Cycle 24 accumulated in 2008 ranged from zero to unprecedented levels of solar activity. The predictions formed an almost normal distribution, centered on the average amplitude of all preceding solar cycles. The average of the current compilation of 105 predictions of the annual-average sunspot number is 106 +/- 31, slightly lower than earlier compilations but still with a wide distribution. Solar Cycle 24 is on track to have a below-average amplitude, peaking at an annual sunspot number of about 80. Our need for solar activity predictions and our desire for those predictions to be made ever earlier in the preceding solar cycle will be discussed. Solar Cycle 24 has been a below-average sunspot cycle. There were peaks in the daily and monthly averaged sunspot number in the Northern Hemisphere in 2011 and in the Southern Hemisphere in 2014. With the rapid increase in solar data and capability of numerical models of the solar convection zone we are developing the ability to forecast the level of the next sunspot cycle. But predictions based only on the statistics of the sunspot number are not adequate for predicting the next solar maximum. I will describe how we did in predicting the amplitude of Solar Cycle 24 and describe how solar polar field predictions could be made more accurate in the future.
Precipitation assimilation poses a special challenge in that the forward model for rain in a global forecast system is based on parameterized physics, which can have large systematic errors that must be rectified to use precipitation data effectively within a standard statistical analysis framework. We examine some key issues in precipitation assimilation and describe several exploratory studies in assimilating rainfall and latent heating information in NASA's global data assimilation systems using the forecast model as a weak constraint. We present results from two research activities. The first is the assimilation of surface rainfall data using a time-continuous variational assimilation based on a column model of the full moist physics. The second is the assimilation of convective and stratiform latent heating retrievals from microwave sensors using a variational technique with physical parameters in the moist physics schemes as a control variable. We will show the impact of assimilating these data on analyses and forecasts. Among the lessons learned are (1) that the time-continuous application of moisture/temperature tendency corrections to mitigate model deficiencies offers an effective strategy for assimilating precipitation information, and (2) that the model prognostic variables must be allowed to directly respond to an improved rain and latent heating field within an analysis cycle to reap the full benefit of assimilating precipitation information. of microwave radiances versus retrieval information in raining areas, and initial efforts in developing ensemble techniques such as Kalman filter/smoother for precipitation assimilation. Looking to the future, we discuss new research directions including the assimilation
Improvements to global and regional numerical weather prediction have been demonstrated through assimilation of data from NASA's Atmospheric Infrared Sounder (AIRS). Current operational data assimilation systems use AIRS radiances, but impact on regional forecasts has been much smaller than for global forecasts. Previously, it has been shown that cloud top designation associated with quality control procedures within the Gridpoint Statistical Interpolation (GSI) system used operationally by a number of Joint Center for Satellite Data Assimilation (JCSDA) partners may not provide the best representation of cloud top pressure (CTP). Because this designated CTP determines which channels are cloud-free and, thus, available for assimilation, ensuring the most accurate representation of this value is imperative to obtaining the greatest impact from satellite radiances. This paper examines the assimilation of hyperspectral sounder data used in operational numerical weather prediction by comparing analysis increments and numerical forecasts generated using operational techniques with a research technique that swaps CTP from the Moderate-resolution Imaging Spectroradiometer (MODIS) for the value of CTP calculated from the radiances within GSI.