Search NASA⌕ Search

SEARCH · Search NASA

Results for “forecast 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 505 records · Page 28

Solar Activity Modeling: From Subgranular Dynamical Scales to the Solar Cycles

The dynamical effects of solar magnetoconvection span a wide range spatial and temporal scales that extend from the interior to the corona and from fast turbulent motions to global magnetic activity. To study the solar activity on short temporal scales (from minutes to hours), we use 3D radiative MHD simulations that allow us to investigate complex turbulent interactions that drive various phenomena, such as plasma eruptions, spontaneous formation of magnetic structures, funnel-like structures and magnetic loops in the corona, and others. In particular, we focus on multi-scale processes of energy exchange across layers of the solar interior and atmosphere, which contribute to coronal heating and eruptive dynamics. For modeling global-scale activity, we use a data assimilation approach that has demonstrated great potential for building reliable long-term forecasts of solar activity. In particular, it has been shown that the Ensemble Kalman Filter (EnKF) method applied to the Parker-Kleeorin-Ruzmakin dynamo model is capable of predicting solar activity up to one sunspot cycle ahead in time, as well as estimating the properties of the next cycle a few years before it begins. In this presentation, using the available magnetogram data, we discuss development of the methodology and forecast quality criteria (including forecast uncertainties and sources of errors). We demonstrate the influence of observational limitations on prediction accuracy, and we present the EnKF predictions of the upcoming Solar Cycle (25) based on both the sunspot number series and observed magnetic fields and discuss the uncertainties and potential of the data assimilation approach for modeling and forecasting solar activity.

Kitiashvili, I. N.↗

Optimum employment of satellite indirect soundings as numerical model input

The characteristics of satellite-derived temperature soundings that would significantly affect their use as input for numerical weather prediction models were examined. Independent evaluations of satellite soundings were emphasized to better define error characteristics. Results of a Nimbus-6 sounding study reveal an underestimation of the strength of synoptic scale troughs and ridges, and associated gradients in isobaric height and temperature fields. The most significant errors occurred near the Earth's surface and the tropopause. Soundings from the TIROS-N and NOAA-6 satellites were also evaluated. Results again showed an underestimation of upper level trough amplitudes leading to weaker thermal gradient depictions in satellite-only fields. These errors show a definite correlation to the synoptic flow patterns. In a satellite-only analysis used to initialize a numerical model forecast, it was found that these synoptically correlated errors were retained in the forecast sequence.

Horn, L. H.↗

Modeling the effects of dust-radiative forcing on the movement of Hurricane Helene (2006)

This study is the first to relate Saharan dust and tropical cyclone (TC) movement over the eastern and central Atlantic Ocean (ECAO) using observations and modeling. An eastward position forecast bias for TCs over ECAO has been noticed and is potentially related to the neglect of dust-radiative forcing in numerical models. Using 11-year observations we found that aerosol optical depth and TC westward speed over ECAO are correlated (0.62) with a confidence level of 99%. The Weather Research and Forecasting dust model was used to quantitatively study the impact of the dust-radiative forcing on Helene’s (2006) track forecasts with 3 different microphysics schemes. The dust-radiative forcing increases the temperature in the warm Saharan Air Layer, thus increases the temperature gradient and modifies wind speed and/or direction along the southern flank of the dust plume, improving Helene’s track. Including dust-radiative forcing reduces the error in 7-day position forecasts by 27% (~205km).

58 GEOSCIENCES↗

A Systematic Error Correction Method for TOVS Radiances

Treatment of systematic errors is crucial for the successful use of satellite data in a data assimilation system. Systematic errors in TOVS radiance measurements and radiative transfer calculations can be as large or larger than random instrument errors. The usual assumption in data assimilation is that observational errors are unbiased. If biases are not effectively removed prior to assimilation, the impact of satellite data will be lessened and can even be detrimental. Treatment of systematic errors is important for short-term forecast skill as well as the creation of climate data sets. A systematic error correction algorithm has been developed as part of a 1D radiance assimilation. This scheme corrects for spectroscopic errors, errors in the instrument response function, and other biases in the forward radiance calculation for TOVS. Such algorithms are often referred to as tuning of the radiances. The scheme is able to account for the complex, air-mass dependent biases that are seen in the differences between TOVS radiance observations and forward model calculations. We will show results of systematic error correction applied to the NOAA 15 Advanced TOVS as well as its predecessors. We will also discuss the ramifications of inter-instrument bias with a focus on stratospheric measurements.

Joiner, Joanna↗

Transformers and Long Short-Term Memory Transfer Learning for GenIV Reactor Temperature Time Series Forecasting

Automated monitoring of the coolant temperature can enable autonomous operation of generation IV reactors (GenIV), thus reducing their operating and maintenance costs. Automation can be accomplished with machine learning (ML) models trained on historical sensor data. However, the performance of ML usually depends on the availability of large amount of training data, which is difficult to obtain for GenIV, as this technology is still under development. We propose the use of transfer learning (TL), which involves utilizing knowledge across different domains, to compensate for this lack of training data. TL can be used to create pre-trained ML models with data from small-scale research facilities, which can then be fine-tuned to monitor GenIV reactors. In this work, we develop pre-trained Transformer and long short-term memory (LSTM) networks by training them on temperature measurements from thermal hydraulic flow loops operating with water and Galinstan fluids at room temperature at Argonne National Laboratory. The pre-trained models are then fine-tuned and re-trained with minimal additional data to perform predictions of the time series of high temperature measurements obtained from the Engineering Test Unit (ETU) at Kairos Power. The performance of the LSTM and Transformer networks is investigated by varying the size of the lookback window and forecast horizon. The results of this study show that LSTM networks have lower prediction errors than Transformers, but LSTM errors increase more rapidly with increasing lookback window size and forecast horizon compared to the Transformer errors.

LSTM↗

Jet stream winds - Comparisons of analyses with independent aircraft data over Southwest Asia

Cruise-level wind data from commercial aircraft are obtained, and these data are compared with operational jet stream analyses over southwest Asia, an area of limited conventional data. Results from an ensemble of 11 cases during January 1989 and individual cases during December 1988-March 1989 are presented. The key results are: (1) European Centre for Medium-Range Weather Forecasts (ECMWF), National Meteorological Center, and United Kingdom Meteorological Office analyses of the subtropical jet in southwest Asia are 11 percent, 17 percent, and 17 percent weaker, respectively, than aircraft observations; (2) analyzed poleward shears range up to 1 f (0.00007/s) compared with up to 3 f (0.00021/s) in the aircraft observations where f is the local Coriolis parameters; (3) the ECMWF errors are largest at the base of the jet; (4) the mean ECMWF core location is latitudinally correct but has an rms latitude variance of 1.5 deg; (5) isolated erroneous radiosondes produce unmeteorological structures in the analyzed subtropical jet stream; and (6) the increased utilization of automated aircraft reports is likely to produce a spurious secular increase in the apparent strength of the jets. The magnitude and spatial extent of the errors seen are near limits of current GCM resolution (100 km) but should be resolvable. The results imply that studies of GCM systematic jet stream wind errors in weather and climate forecasts must be interpreted with caution in this region.

Tenenbaum, J.↗

Process-oriented evaluation using ARM observations to improve clouds and radiation in a global atmospheric model

The goal of the research is to improve understanding of cloud, radiation and precipitation processes and their representation in numerical models for weather and climate using observations from the ARM/ASR program. Global models need to represent all cloud types across different weather regimes from the tropics to the poles and ARM provides long-term observations at different geographical sites that cover these different regimes. Combined with satellite data to give a global context, the observations are used to evaluate the systematic errors in a global model for predictions of cloud, radiation and precipitation. The focus is on understanding the links to underlying physical processes to pinpoint the sources of error and improve the model representation of these processes. The research covers boundary layer cloud over the continental U.S., marine stratocumulus and shallow cumulus in the subtropics, mixed-phase clouds over the Arctic and Southern Ocean, and high-altitude cirrus ice clouds in the tropics. The model used is the operational global weather prediction system at the European Centre for Medium-range Weather Forecasts (ECMWF). The research has led to numerous developments, several of which have been implemented operationally leading to improvements in everyday forecasts. Many of the systematic errors are common across different modelling systems and the wider implications from the research for other weather and climate models are highlighted.

54 ENVIRONMENTAL SCIENCES↗

Machine Learning–Adjusted WRF Forecasts to Support Wind Energy Needs in Black Start Operations

Abstract The push for increased capacity of renewable sources of electricity has led to the growth of wind-power generation, with a need for accurate forecasts of winds at hub height. Forecasts for these levels were uncommon until recently, and that, combined with the nocturnal collapse of the well-mixed boundary layer and daytime growth of the boundary layer through the levels important for energy generation, has contributed to errors in numerical modeling of wind generation resources. The present study explores several machine learning algorithms to both forecast and correct standard WRF Model forecasts of winds and temperature at hub height within wind turbine plants over several different time periods that are critical for the anticipation of potential blackouts and aiding in black start operations on the power grid. It was found that mean square error for day-2 wind forecasts from the WRF Model can be improved by over 90% with the use of a multioutput neural network, and that 60-min forecasts of WRF error, which can then be used to adjust forecasts, can be made with an LSTM with great accuracy. Nowcasting of temperature and wind speed over a 10-min period using an LSTM produced very low error and especially skillful forecasts of maximum and minimum values over the turbine plant area.

17 WIND ENERGY↗

Why Is Rainfall Error Analysis Requisite for Data Assimilation and Climate Modeling?

Given the large temporal and spatial variability of precipitation processes, errors in rainfall observations are difficult to quantify yet crucial to making effective use of rainfall data for improving atmospheric analysis, weather forecasting, and climate modeling. We highlight the need for developing a quantitative understanding of systematic and random errors in precipitation observations by examining explicit examples of how each type of errors can affect forecasts and analyses in global data assimilation. We characterize the error information needed from the precipitation measurement community and how it may be used to improve data usage within the general framework of analysis techniques, as well as accuracy requirements from the perspective of climate modeling and global data assimilation.

Hou, Arthur Y.↗

Forecasting Day-Ahead Solar Irradiance for Puerto Rico Using the WRF Model and NSRDB

Accurately predicting solar energy resources is a major challenge in integrating photovoltaics generation on the electric grid. Numerical weather prediction has been recognized by the solar energy community as a major approach to provide solar resource forecasts at various locations and for a variety of timescales. In this study, as a part of the Puerto Rico Grid Resilience and Transitions to 100% Renewable Energy Study (PR100), we develop day-head solar irradiance forecast data using the Weather Research and Forecasting (WRF) model at 3 km and hourly/5-minute. The global horizontal irradiance (GHI) and direct normal irradiance (DNI) forecasts simulated from the WRF model are postprocessed by a simple optimization method using satellite-derived gridded observations from the National Solar Radiation Data Base (NSRDB) to reduce error and bias of the solar irradiance forecasts covering 2018-2020. The NSRDB contributes to improving the GHI and DNI forecasts and also offers the opportunity for an in-depth analysis to evaluate their accuracy over a wide range of Puerto Rico regions. Preliminary results show overall improvements of GHI forecasts up to 37% (DNI: 15%) for mean absolute error and 97% (DNI: 76%) for mean bias error by applying a postprocessing technique to WRF model output.

data models↗

A hybrid wavelet de-noising and Rank-Set Pair Analysis approach for forecasting hydro-meteorological time series

Highlights: • A new hybrid approach is proposed to improve forecasts of hydro-meteorological time series. • Rank-Set Pair Analysis combined with wavelet de-noising markedly improves forecasting accuracy. • The performance of the proposed approach proves best among its present competitors even when the extreme value occurs. Accurate, fast forecasting of hydro-meteorological time series is presently a major challenge in drought and flood mitigation. This paper proposes a hybrid approach, wavelet de-noising (WD) and Rank-Set Pair Analysis (RSPA), that takes full advantage of a combination of the two approaches to improve forecasts of hydro-meteorological time series. WD allows decomposition and reconstruction of a time series by the wavelet transform, and hence separation of the noise from the original series. RSPA, a more reliable and efficient version of Set Pair Analysis, is integrated with WD to form the hybrid WD-RSPA approach. Two types of hydro-meteorological data sets with different characteristics and different levels of human influences at some representative stations are used to illustrate the WD-RSPA approach. The approach is also compared to three other generic methods: the conventional Auto Regressive Integrated Moving Average (ARIMA) method, Artificial Neural Networks (ANNs) (BP-error Back Propagation, MLP-Multilayer Perceptron and RBF-Radial Basis Function), and RSPA alone. Nine error metrics are used to evaluate the model performance. Compared to three other generic methods, the results generated by WD-REPA model presented invariably smaller error measures which means the forecasting capability of the WD-REPA model is better than other models. The results show that WD-RSPA is accurate, feasible, and effective. In particular, WD-RSPA is found to be the best among the various generic methods compared in this paper, even when the extreme events are included within a time series.

58 GEOSCIENCES↗

Observation Impact and Information Retention in the Lower Troposphere of the GMAO GEOS Data Assimilation System

In this study, we have assessed the effectiveness of the use of existing observing systems in the lower troposphere in the GEOS hybrid–4DEnVar data assimilation system through a set of observing system experiments. The results show that microwave radiances have a large impact in the Southern Hemisphere and Tropical ocean, but the large influence is mostly observed above 925 hPa and dissipates relatively quickly with longer forecast lead times. Conventional data information holds better in the forecast ranging from the surface to 100 hPa, depending on the field evaluated, in the Northern Hemisphere and lowest model levels in the Tropics. Infrared radiances collectively have much less impact in the lower troposphere. Removing surface observations has small but persistent impact on specific humidity in the upper atmosphere, but small or negligible impact on planetary boundary layer (PBL) height and temperature. The model responses to the incremental analysis update (IAU) forcing are also analyzed. In the IAU assimilation window, the physics responds strongly to the IAU forcing in the lower troposphere, and the changes of physics tendency in the lower troposphere and hydrodynamics tendency in the mid- and upper troposphere are viewed as beneficial to the reduction of state error covariance. In the subsequent forecast, the model tendencies continue to deviate further from the original free forecast with forecast lead times around 300–400 hPa, but physics tendency has showed signs of returning to its original free forecast mechanisms at 1-day forecast in the lower troposphere.

GEOS↗

A cross-dimensional analysis of data-driven short-term load forecasting methods with large-scale smart meter data

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.

Li, Han↗

Short-range forecasts with the GISS model of the global atmosphere

A nine-layer, primitive equation (PE) model of the global atmosphere developed at the Goddard Institute for Space Studies (GISS) has been used to generate six 48-hr forecasts during December 1972 and January 1973. Operational analyses north of 18 N and experimental global analyses made available by the National Meteorological Center (NMC), NOAA, were used as the initial conditions; the operational analyses were used to verify the forecasts at 12-hr intervals over the Northern Hemisphere north of 22 N. The combined analyses were used to verify the forecasts in the global domain. RMS errors of the sea-level pressure, 1000-mb heights, and vector geostrophic winds, and 500-mb heights and vector geostrophic winds indicate that the GISS forecasts have skill comparable to those made by operational PE models.

Druyan, L. M.↗

Dynamically stratified Monte Carlo forecasting

A new method for performing Monte Carlo forecasts is introduced. The method, called dynamic stratification, selects initial perturbations based on a stratification of the error distribution. A simple implementation is presented in which the error distribution used for the stratification is estimated from a linear model derived from a large ensemble of 12-h forecasts with the full dynamic model. The stratification thus obtained is used to choose a small subsample of initial states with which to perform the dynamical Monte Carlo forecasts. Several test cases are studied using a simple two-level general circulation model with uncertain initial conditions. It is found that the method provides substantial reductions in the sampling error of the forecast mean and variance when compared to the more traditional approach of choosing the initial perturbations at random. The degree of improvement, however, is sensitive to the nature of the initial error distribution and to the base state. In practice the method may be viable only if the computational burden involved in obtaining an adequate estimate of the error distribution is shared with the data-assimilation procedure.

Schubert, Siegfried↗

Development of a climatological data base to help forecast cloud cover conditions for shuttle landings at the Kennedy Space Center

The Space Shuttle is an extremely weather sensitive vehicle with very restrictive constraints for both launches and landings. The most important difference between Shuttle and normal aircraft landings is that the Shuttle has no go-around capability once it begins its decent into the earth's atmosphere. The de-orbit burn decision is generally made approximately 90 minutes before landing requiring a forecast with little room for error. Because of the Shuttle's rapid re-entry to earth, the pilot must be able to see all runway and visual navigation aids from high altitude to land the Shuttle. In addition, the heat resistant tiles which are used to protect the Shuttle during its re-entry into the earth's atmosphere are extremely sensitive to any type of precipitation. Extensive damage to these tiles could occur if the Shuttle passes through any cloud that contains precipitation size particles. To help guard against changing weather conditions or any type of weather problems that might occur prior to landing, flight rules have been developed as guidelines for all landings. Although the rules vary depending on the location of the landing (Kennedy Space Center or Edwards AFB), length of mission, and weight of vehicle, most of the rules can be condensed into 4 major groupings. These are: (1) Cloud ceilings should not be less than 3048 m (10,000 feet), (2) Visibility should not be less than 13 km (7 nm), (3) Cross-wind no greater than 5-8 m/s (10-15 knots); and (4) No showers or thunderstorms at or within 56 km (30 nm) of the Shuttle Landing Facility. This study consisted of developing a climatological database of the Shuttle Landing Facility (SLF) surface observations and performing an analysis of observed conditions one and two hours subsequent to given conditions at the SLF to help analyze the 0.2 cloud cover rule. Particular emphasis was placed on Shuttle landing weather violations and the amounts of cloud cover below 3048 m (10,000 ft.). This analysis has helped to determine the best and worst times to land the Shuttle at KSC. In addition, nomograms have been developed to help forecasters make cloud cover forecasts for End of Mission (EOM) and Return to Launch Site (RTLS) at KSC. Results of categorizing this data by month, season, time of day, and surface and upper-air wind direction are presented.

Atchison, M. Kevin↗

Modeling the 4D discharge of lithium-ion batteries with a multiscale time-dependent deep learning framework

The lithium-ion battery (LIB) field is moving towards the direction of investigating spatially resolved physical phenomena in the 3D porous microstructure of electrodes. These pore-scale simulations give new insights into the local dynamics of lithiation/de-lithiation and charge transport, Nevertheless, the computational time of these simulations limits the integration of these models in optimization workflows of cycling conditions or electrode manufacturing processes. Machine learning models present a way of assessing in real-time the performance of materials. While several successful techniques for replicating simulations with machine learning have been proposed, this case study presents a more demanding problem, due to the necessity of understanding the behavior of heterogeneous 3D local data, as it evolves in time: this poses both a scientific and a technical challenge. To this end, we propose an autoregressive multiscale convolutional neural network model to predict relevant quantities at the pore-scale in the solid phase: the lithium concentration (in the active material) and potential (in the active material and carbon binder). Here, these are ultimately used to reconstruct the battery discharge curve. 3D images of the electrode microstructures are the input to the network, trained with a dataset of finite element method simulations to predict the discharge behavior of the cathode side in lithium ion batteries. We propose this machine learning model as a proof-of-concept of the applicability of multiscale networks for time-dependent physics problems. The trained model exhibits very high accuracy (with errors lower than 2 %) in forecasting the discharge behavior of new unseen cathodes.

25 ENERGY STORAGE↗