Search NASA⌕ Search

SEARCH · Search NASA

Results for “FORECASTING MODEL”

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 865 records · Page 48
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Improving Seasonal Forecast Using Probabilistic Deep Learning

The path toward realizing the potential of seasonal forecasting and its socioeconomic benefits relies on improving general circulation model (GCM) based dynamical forecast systems. To improve dynamical seasonal forecasts, it is crucial to set up forecast benchmarks, and clarify forecast limitations posed by model initialization errors, formulation deficiencies, and internal climate variability. With huge costs in generating large forecast ensembles, and limited observations for forecast verification, the seasonal forecast benchmarking and diagnosing task proves challenging. Here, we develop a probabilistic deep learning-based statistical forecast methodology, drawing on a wealth of climate simulations to enhance seasonal forecast capability and forecast diagnosis. By explicitly modeling the internal climate variability and GCM formulation differences, the proposed Conditional Generative Forecasting (CGF) methodology enables bypassing crucial barriers in dynamical forecast, and offers a top-down viewpoint to examine how complicated GCMs encode the seasonal predictability information. We apply the CGF methodology for global seasonal forecast of precipitation and 2 m air temperature, based on a unique data set consisting 52,201 years of climate simulation. Results show that the CGF methodology can faithfully represent the seasonal predictability information encoded in GCMs. We successfully apply this learned relationship in real-world seasonal forecast, achieving competitive performance compared to dynamical forecasts. Using this CGF as benchmark, we reveal the impact of insufficient forecast spread sampling that limits the skill of the considered dynamical forecast system. Finally, we introduce different strategies for composing ensembles using the CGF methodology, highlighting the potential for leveraging the strengths of multiple GCMs to achieve advantgeous seasonal forecast.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Deep Learning for Ensemble Forecasting

Focal Area: (2) Predictive modeling through the use of AI techniques and AI-derived model components and the use of AI and other tools to design a prediction system comprising a hierarchy of models. Science Challenge: While both climate and weather forecast systems have continued to improve due to substantial efforts to improve computational capabilities, observations, and numerical models, the atmosphere is a chaotic system, and this puts a fundamental limit on our ability to make predictions. Forecasts made by high-resolution models initialized with only slightly different atmospheric states can quickly diverge. Quantifying uncertainty in forecasts is essential to adequately understand them and to make the best-informed policy decisions particularly when it comes to hydrology, extreme weather (including extreme precipitation events), and climate.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Upper Troposphere Smoke Injection From Large Areal Fires

Abstract Large areal fires, such as those ignited following a nuclear detonation, can inject smoke into the upper troposphere and lower stratosphere. Detailed fire simulations allow for assessment of how local weather interacts with these fires and affects smoke lofting. In this study, we employ the fire simulation package in the Weather Research and Forecasting (WRF‐Fire) model, Version 4.0.1, to explore how smoke lofting from a fire burning a homogeneous fuel bed changes with varying local winds, relative humidity, and atmospheric boundary‐layer stability for two different‐sized areal fires. The presence of moisture has the greatest influence on the results by raising the altitude of lofting, while faster wind speeds dampen lofting and lower the injection height. Stably stratified conditions inhibit plume propagation compared with neutrally stratified conditions, although the impact of stability is not as strong as that of moisture and winds. These findings highlight the importance of using an appropriate atmospheric profile when simulating large fires, as the local weather can have a meaningful influence on smoke lofting.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

A Statistical Interpolation Code for Ocean Analysis and Forecasting

Abstract We present a data assimilation package for use with ocean circulation models in analysis, forecasting, and system evaluation applications. The basic functionality of the package is centered on a multivariate linear statistical estimation for a given predicted/background ocean state, observations, and error statistics. Novel features of the package include support for multiple covariance models, and the solution of the least squares normal equations either using the covariance matrix or its inverse—the information matrix. The main focus of this paper, however, is on the solution of the analysis equations using the information matrix, which offers several advantages for solving large problems efficiently. Details of the parameterization of the inverse covariance using Markov random fields are provided and its relationship to finite-difference discretizations of diffusion equations are pointed out. The package can assimilate a variety of observation types from both remote sensing and in situ platforms. The performance of the data assimilation methodology implemented in the package is demonstrated with a yearlong global ocean hindcast with a 1/4° ocean model. The code is implemented in modern Fortran, supports distributed memory, shared memory, multicore architectures, and uses climate and forecasts compliant Network Common Data Form for input/output. The package is freely available with an open source license from www.tendral.com/tsis/ .

Srinivasan, Ashwanth↗
DOE OSTI2023

The Potential Benefits of Handling Mixture Statistics via a Bi-Gaussian EnKF: Tests With All-Sky Satellite Infrared Radiances

The meteorological characteristics of cloudy atmospheric columns can be very different from their clear counterparts. Thus, when a forecast ensemble is uncertain about the presence/absence of clouds at a specific atmospheric column (i.e., some members are clear while others are cloudy), that column's ensemble statistics will contain a mixture of clear and cloudy statistics. Such mixtures are inconsistent with the ensemble data assimilation algorithms currently used in numerical weather prediction. Hence, ensemble data assimilation algorithms that can handle such mixtures can potentially outperform currently used algorithms. In this study, we demonstrate the potential benefits of addressing such mixtures through a bi-Gaussian extension of the ensemble Kalman filter (BGEnKF). The BGEnKF is compared against the commonly used ensemble Kalman filter (EnKF) using perfect model observing system simulated experiments (OSSEs) with a realistic weather model (the Weather Research and Forecast model). Synthetic all-sky infrared radiance observations are assimilated in this study. In these OSSEs, the BGEnKF outperforms the EnKF in terms of the horizontal wind components, temperature, specific humidity, and simulated upper tropospheric water vapor channel infrared brightness temperatures. This study is one of the first to demonstrate the potential of a Gaussian mixture model EnKF with a realistic weather model. Our results thus motivate future research toward improving numerical Earth system predictions though explicitly handling mixture statistics.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Tracing and Forecasting Metabolic Indices of Cancer Patients Using Patient-Specific Deep Learning Models

We develop a patient-specific dynamical system model from the time series data of the cancer patient’s metabolic panel taken during the period of cancer treatment and recovery. The model consists of a pair of stacked long short-term memory (LSTM) recurrent neural networks and a fully connected neural network in each unit. It is intended to be used by physicians to trace back and look forward at the patient’s metabolic indices, to identify potential adverse events, and to make short-term predictions. When the model is used in making short-term predictions, the relative error in every index is less than 10% in the L ∞ norm and less than 6.3% in the L 1 norm in the validation process. Once a master model is built, the patient-specific model can be calibrated through transfer learning. As an example, we obtain patient-specific models for four more cancer patients through transfer learning, which all exhibit reduced training time and a comparable level of accuracy. This study demonstrates that this modeling approach is reliable and can deliver clinically acceptable physiological models for tracking and forecasting patients’ metabolic indices.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Distribution Substation Planning Toolkit (dsp-toolkit) v1.0

The Distribution Substation Planning Toolkit (DSP Toolkit) is a software suite designed to streamline the planning and optimization of distribution substations. This toolkit offers a comprehensive set of tools and APIs for data curation, short-term electric load forecasting, and weather-sensitive load adjustment, making it an essential resource for utility companies, engineers, and researchers. Features • Data Preprocessing and Curation: Efficiently manage and preprocess large datasets to ensure high-quality input for analysis. • Short-Term Load Forecasting: Utilize data-driven models to predict short-term electric loads accurately. • Weather-Sensitive Modeling: Automatically adjust load forecasts based on weather data to predict future peak demands more precisely. Uses The DSP Toolkit is ideal for planning and optimizing distribution substations, providing a user-friendly interface and comprehensive documentation. It is suitable for both novice and experienced users, facilitating efficient and accurate planning processes. Advantages • Efficiency: Automates complex planning tasks, reducing manual effort and minimizing errors. • Scalability: Handles large datasets and complex models, making it suitable for large-scale projects. • Community and Support: Open-source with active community contributions, ensuring continuous improvement and support. • Extensibility: Easily extendable with custom modules and plugins, allowing users to tailor the toolkit to their specific needs. The DSP Toolkit stands out by offering a robust, flexible, and user-friendly solution for distribution substation planning. Public Abstract

Li, Han [Lawrence Berkeley National Laboratory (LB↗
DOE OSTI2025

Short-Term Forecasting of Thermostatic and Residential Loads Using Long Short-Term Memory Recurrent Neural Networks

Internet of Things (IoT) devices in smart grids enable intelligent energy management for grid managers and personalized energy services for consumers. Investigating a smart grid with IoT devices requires a simulation framework with IoT devices modeling. However, there lack comprehensive study on the modeling of IoT devices in smart grids. This paper investigates the IoT device modeling of a thermostatic load and implements the recurrent neural networks model for short-term load forecasting in this IoT-based thermostatic load. The recurrent neural network structure is leveraged to build a load forecasting model on temporal correlation. The temporal recurrent neural network layers including long short-term memory cells are employed to learn the data from both the simulation platform and New South Wales residential datasets. The simulation results are provided for demonstration.

electric load forecasting↗
DOE OSTI2017

Investigation of the N 2 O emission strength in the U. S. Corn Belt

Nitrous oxide (N 2 O) has a high global warming potential and depletes stratospheric ozone. The U. S. Corn Belt plays an important role in the global anthropogenic N 2 O budget. To date, studies on local surface N 2 O emissions and the atmospheric N 2 O budget have commonly used Lagrangian models. Here, in the present study, we used an Eulerian model - Weather Research and Forecasting Chemistry (WRF-Chem) model to investigate the relationships between N 2 O emissions in the Corn Belt and observed atmospheric N 2 O mixing ratios. We derived a simple equation to relate the emission strengths to atmospheric N 2 O mixing ratios, and used the derived equation and hourly atmospheric N 2 O measurements at the KCMP tall tower in Minnesota to constrain agricultural N 2 O emissions. The modeled spatial patterns of atmospheric N 2 O were evaluated against discrete observations at multiple tall towers in the NOAA flask network. After optimization of the surface flux, the model reproduced reasonably well the hourly N 2 O mixing ratios monitored at the KCMP tower. Agricultural N 2 O emissions in the EDGAR42 database needed to be scaled up by 19.0 to 28.1 fold to represent the true emissions in the Corn Belt for June 1–20, 2010 - a peak emission period. Optimized mean N 2 O emissions were 3.00–4.38, 1.52–2.08, 0.61–0.81 and 0.56–0.75 nmol m -2 s -1 for June 1–20, August 1–20, October 1–20 and December 1–20, 2010, respectively. The simulated spatial patterns of atmospheric N 2 O mixing ratios after optimization were in good agreement with the NOAA discrete observations during the strong emission peak in June. Such spatial patterns suggest that the underestimate of emissions using IPCC (Inter-governmental Panel on Climate Change) inventory methodology is not dependent on tower measurement location.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Tracking Aerosol Convection Interactions Experiment (TRACER) Field Campaign Report

Convective clouds serve a critical role in the Earth’s energy and water cycles through their transport of heat, moisture, momentum, and chemical species through the troposphere driving the global circulation (e.g., Hartmann et al. 1984, Del Genio et al. 2012, Su et al. 2014). On more local scales, convective clouds impact the atmospheric heating profile through diabatic heating effects, removal of water from the atmospheric column through precipitation, and conditioning of the local environment impacting further development of clouds (e.g., Sullivan and Voigt 2021). These critical roles underscore the importance of realistic representation of convective processes across scales of models from large-eddy simulation (LES), to convection-permitting models (CPM; e.g., Kendon et al. 2020, Marinescu et al. 2021), to numerical weather prediction (NWP) models used for operational weather forecasting, to Earth system models used to predict climate sensitivity (Sanderson et al. 2011, Sherwood et al. 2014, Tomassini et al. 2014, Zhao et al. 2016, Cronin et al. 2017). A key component of improving model representation of convective clouds is better quantification and parameterization of updraft microphysics and dynamics, including their interactions with the surrounding environment and storm organization (Bony et al. 2015, Hagos and Houze 2016, Donner et al. 2016, Morrison et al. 2020). Aerosol is an important environmental factor that could affect convective clouds and precipitation since cloud droplet and ice formation processes are initiated by it. Andrae et al. (2004) hypothesized that aerosols associated with increased biomass burning particles acting as cloud condensation nuclei (CCN) result in smaller and more monodisperse cloud droplets leading to suppression of warm rain formation, ultimately leading to more cloud water being lofted above the freezing level based on observations in the Amazon region. The subsequent increase in latent heat release increases the buoyancy of rising convective parcels invigorating the deep convection. This work was followed by a description of the theoretical basis for this “cold-phase invigoration” by Rosenfeld et al. (2008), who argued that it could have a significant effect for deep convective clouds with warm cloud-bases. Several modeling studies (e.g., Khain et al. 2005, 2009, van den Heever et al. 2006, Fan et al. 2007, 2009, 2012, Lee et al. 2008, Storer et al. 2010, Lebo et al. 2012, Storer and van den Heever 2013, Chen et al. 2020, Dagan et al. 2022) have investigated these aerosol-convection interactions and the environmental factors that influence their relative importance and magnitude. More recently, several studies have indicated that “warm-phase invigoration”, the enhancement of convection through condensational heating, also appears to play a role in enhancing both shallow cumuli (Seiki and Nakajima 2014, Saleeby et al 2015) and deeper tropical convection (Lebo and Seinfeld 2011, Khain et al. 2012, Sheffield et al 2015, Fan et al. 2018, Igel and van den Heever 2021), as well as Houston thunderstorms (Fan et al. 2007, 2020). However, still other studies have provided additional evidence of systematic biases in simulated convective outflow ice size distribution properties, which are consistent with a lack of poorly understood secondary ice production within convective updrafts (e.g., Fridlind et al. 2017). To help address these critical gaps in our understanding of cloud processes, aerosol processes and aerosol-cloud interactions, the Tracking Aerosol Convection Interactions Experiment was designed building upon efforts by the Aerosol, Cloud, Precipitation and Climate (ACPC) Initiative (http://acpcintiative.org/), a joint effort of the International Geosphere-Biosphere Programme (IGBP) and the World Climate Research Program (WCRP) that focused on resolving uncertainties in the interactions between aerosol and clouds towards better understanding the role that these interactions play in the climate system. The TRACER campaign was motivated by recommendations from a number of pilot studies undertaken by ACPC (van den Heever et al. 2017, Fridlind et al. 2019, Hu et al. 2019, Fan et al. 2020, Marinescu et al. 2021, Hernandez-Deckers et al. 2022) that pointed towards the southeastern Texas region as a locale where aerosol-convection interactions could be studied owing to the copious occurrence of isolated convection during the summer months accompanied by diverse and significant sources of aerosols from both anthropogenic and natural sources. The TRACER campaign began on 01 October 2021 and extended through 30 September 2022 with an intensive operational period (IOP) during June-September 2022. Three main sites (Table 1) were managed by the U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) user facility.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Assessing Impacts of Waves on Hub-Height Winds off the U.S. West Coast Using Lidar Buoys and Coupled Modeling Approaches

Given the importance of offshore wind energy development to the U.S. clean energy targets, it is vital to be able to characterize the wind resource in that environment accurately. Toward that end, two Bureau of Ocean Energy Management buoys equipped with Doppler lidar are being maintained by Pacific Northwest National Laboratory on behalf of the Department of Energy and deployed to regions of potential offshore wind development. In addition to standard meteorological and oceanographic measurements, the buoys document the wind profile between about 40 m and 250 m above the sea surface through Doppler lidar retrievals. After a multiyear deployment of two buoys along the U.S. East Coast, the buoys were redeployed to the U.S. West coast from 2020 – 2022 to locations near the Humboldt and Morro Bay lease areas. The buoys provide nearly continuous, multiyear datasets that can be used to evaluate predictions of hub-height (~100 m) wind speed for standard atmospheric models in the region. In the absence of measurements at the study site, offshore wind developers rely on model-based data to assess site conditions. Potential sources of model error in this environment include under-resolution or misrepresentation of coastal topographically forced flows, marine boundary layer dynamics and the evolution of their associated cloud and turbulence fields, the role of upwelling and other currents on surface heat fluxes into the boundary layer, and the impact of wave fields on surface momentum fluxes and thus the wind speed profile. In particular, most predictive models of wind speed do not predict wave fields at all, relying on parameterizations to represent their effects. In thus study, we focus on evaluating the role of wind / wave interactions on modeled hub-height wind speed and error by using the Coupled Ocean–Atmosphere–Wave–Sediment–Transport Modeling System to capture two-way interactions between an atmospheric model (Weather Research and Forecasting (WRF)) and a wave model (WAVEWATCHIII (WW3)) and compare to both stand-alone WRF and one-way coupled WRF / WW3 configurations. Our approach is similar to that used in Gaudet et al. (2022) to evaluate wind / wave coupling over the U.S. East Coast, but applied to the very different environment of the U.S. West Coast. We show examples for two cases, a cold-season frontal case and a warm-season low-level jet case. We find that wind / wave coupling makes little impact on model error for these cases at the location of the lidar buoys, for which other misrepresentations of model physics seems to be responsible for model-observation discrepancies. However, domain-wide evaluations, which also make use of the National Buoy Data Center network, show that a two-way coupling approach is less prone to systematic errors in the hub-height wind field than the one-way coupled approach. WRF resolution of kilometer-scale or less is needed to properly capture the sharp wind speed gradients that can be found along the coastline, and WW3 simulations driven by the downscaled WRF produce better bulk and spectral wave fields when compared to observations. Implications of the results for wind resource characterization are then discussed.

17 WIND ENERGY↗
DOE OSTI2016

Relating atmospheric N 2 O concentration to N 2 O emission strength in the U. S. Corn Belt

Nitrous oxide (N 2 O) has a high global warming potential and depletes stratospheric ozone. The U. S. Corn Belt plays an important role in the global anthropogenic N 2 O budget. To date, studies on local surface N 2 O emission and the atmospheric N 2 O budget have commonly used Lagrangian models. In the present study, we used an Eulerian model – Weather Research and Forecasting Chemistry (WRF-Chem) model to investigate the relationships between N 2 O emission in the Corn Belt and observed atmospheric N 2 O mixing ratios. Modeled hourly N 2 O mixing ratios were combined with continuous atmospheric N 2 O measurements at the KCMP tall tower in Minnesota to constrain agricultural N 2 O emissions. The modeled spatial patterns of atmospheric N 2 O were validated against discrete observations at multiple tall towers in the NOAA flask network. After optimization of the surface flux, the model reproduced reasonably well the hourly N 2 O mixing ratios monitored at the KCMP tower. Agricultural N 2 O emissions in the EDGAR42 database needed to be scaled up by 19.0 to 28.1 fold to represent the true emission in the Corn Belt from June 1–20, 2010 – a peak emission period. Optimized total N 2 O emissions were 3.00–4.38, 1.52–2.08, 0.61–0.81 and 0.56–0.75 nmol m –2 s –1 from June 1–20, August 1–20, October 1–20 and December 1–20, 2010, respectively. The simulated spatial patterns of atmospheric N 2 O mixing ratios were in good agreement with the NOAA discrete observations during the strong emission peak in June. Such spatial patterns illustrate that the IPCC (Inter-governmental Panel on Climate Change) underestimate of emissions is not dependent on tower measurement location.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Grey-box and ANN-based building models for multistep-ahead prediction of indoor temperature to implement model predictive control

Model-based predictive control (MPC) strategies for heating, ventilation, and air-conditioning (HVAC) systems present an opportunity to lower building energy consumption and operational costs. Such approaches rely on the development of a model to precisely forecast building thermal dynamics, such as room air temperature or heating/cooling rate, and make control-related decisions. The control-oriented modeling of building energy systems should be accurate in predicting indoor conditions and present low computational complexity. These features are the key challenge of implementing advanced control methods such as MPC. Extant studies on building modeling for MPC have focused on step-ahead forecasting techniques to forecast building thermal dynamics, while multistep-ahead forecasting is essential. Moreover, machine learning model suitable in case of the domain-based engineering expertise are also not available. To this aim, we perform a comparative analysis of the grey-box model based on a resistance-capacitance (RC) thermal network and a machine learning model composed of an artificial neural network (ANN) for multistep-ahead prediction of building thermal dynamics using current and historical data. Actual experimental data obtained from the Flexible Research Platform (FRP) in Oak Ridge National Laboratory (US) are used for estimation and validation purposes. The average root mean squared error (RMSE) of the grey-box and ANN models are 0.89 °C and 1.02°C, respectively. Finally, the results indicate that the grey-box model outperforms the ANN model in the considered validation periods in terms of accuracy and prediction stability.

42 ENGINEERING↗
DOE OSTI2016

Predictive Big Data Analytics: A Study of Parkinson’s Disease Using Large, Complex, Heterogeneous, Incongruent, Multi-Source and Incomplete Observations

Background A unique archive of Big Data on Parkinson’s Disease is collected, managed and disseminated by the Parkinson’s Progression Markers Initiative (PPMI). The integration of such complex and heterogeneous Big Data from multiple sources offers unparalleled opportunities to study the early stages of prevalent neurodegenerative processes, track their progression and quickly identify the efficacies of alternative treatments. Many previous human and animal studies have examined the relationship of Parkinson’s disease (PD) risk to trauma, genetics, environment, co-morbidities, or life style. The defining characteristics of Big Data–large size, incongruency, incompleteness, complexity, multiplicity of scales, and heterogeneity of information-generating sources–all pose challenges to the classical techniques for data management, processing, visualization and interpretation. We propose, implement, test and validate complementary model-based and model-free approaches for PD classification and prediction. To explore PD risk using Big Data methodology, we jointly processed complex PPMI imaging, genetics, clinical and demographic data. Methods and Findings Collective representation of the multi-source data facilitates the aggregation and harmonization of complex data elements. This enables joint modeling of the complete data, leading to the development of Big Data analytics, predictive synthesis, and statistical validation. Using heterogeneous PPMI data, we developed a comprehensive protocol for end-to-end data characterization, manipulation, processing, cleaning, analysis and validation. Specifically, we (i) introduce methods for rebalancing imbalanced cohorts, (ii) utilize a wide spectrum of classification methods to generate consistent and powerful phenotypic predictions, and (iii) generate reproducible machine-learning based classification that enables the reporting of model parameters and diagnostic forecasting based on new data. We evaluated several complementary model-based predictive approaches, which failed to generate accurate and reliable diagnostic predictions. However, the results of several machine-learning based classification methods indicated significant power to predict Parkinson’s disease in the PPMI subjects (consistent accuracy, sensitivity, and specificity exceeding 96%, confirmed using statistical n-fold cross-validation). Clinical (e.g., Unified Parkinson's Disease Rating Scale (UPDRS) scores), demographic (e.g., age), genetics (e.g., rs34637584, chr12), and derived neuroimaging biomarker (e.g., cerebellum shape index) data all contributed to the predictive analytics and diagnostic forecasting. Conclusions Model-free Big Data machine learning-based classification methods (e.g., adaptive boosting, support vector machines) can outperform model-based techniques in terms of predictive precision and reliability (e.g., forecasting patient diagnosis). We observed that statistical rebalancing of cohort sizes yields better discrimination of group differences, specifically for predictive analytics based on heterogeneous and incomplete PPMI data. UPDRS scores play a critical role in predicting diagnosis, which is expected based on the clinical definition of Parkinson’s disease. Even without longitudinal UPDRS data, however, the accuracy of model-free machine learning based classification is over 80%. The methods, software and protocols developed here are openly shared and can be employed to study other neurodegenerative disorders (e.g., Alzheimer’s, Huntington’s, amyotrophic lateral sclerosis), as well as for other predictive Big Data analytics applications.

Science & Technology - Other Topics↗
DOE OSTI2022

Deep Learning Provides Substantial Improvements to County‐Level Fire Weather Forecasting Over the Western United States

Abstract The recent wildfires in the western United States during 2018 and 2020 caused record‐breaking fire damage and casualties. Despite remarkable advances in fire modeling and weather forecasting, it remains challenging to anticipate catastrophic wildfire events and associated damage. One key missing component is a fire weather prediction system with sufficiently long lead time capable of providing useful regional details. Here, we develop a hybrid prediction model of wildfire danger called CFS with super resolution (CFS‐SR) as a proof of concept to fill that void. The CFS‐SR model is constructed by integrating the Climate Forecast System version 2 with a deep learning (DL) technique from Single Image Super Resolution, a method widely used in enhancing image resolution. We show that for the 2018–2019 fire season, the CFS‐SR model significantly improves accuracy in forecasting fire weather at lead times of up to 7 days with an enhanced spatial resolution up to 4 km. This level of high resolution provides county‐level fire weather forecast, making it more practical for allocating resources to mitigate wildfire danger. Our study demonstrates that a proper combination of ensemble climate predictions with DL techniques can boost predictability at finer spatial scales, increasing the utility of fire weather forecasts for practical applications.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Recursive Blind Forecasting of Photovoltaic Generation and Consumer Load for Microgrids

Existing forecasting frameworks that predict time-series photovoltaic (PV) generation and consumer load for micro-grids' operation and control assume near-continuous availability of real-time predictors from the field. The incoming data are used to periodically re-train the models and update forecast snapshots over a moving horizon window. However, such frameworks are not resilient to disruptions in data availability caused by losses in communications between the field sensors and data loggers. This paper bridges the shortcoming by leveraging a previously proposed forecasting framework that is resilient to abrupt changes in data quality caused by communication losses. Assuming no availability of real-time field system data, which is typical in extreme weather events such as hurricanes, the framework uses lightweight recursive time-series models to independently forecast solar irradiance, ambient temperature, PV power, and consumer load for three horizon windows: 24 hours, 12 hours, and 1 hour. Four types of ensemble-based regression trees-simple gradient boosted trees (GBR), GBR with an adaptive component (A-GBR), random forests (RF), and extra trees (ExTR)-are leveraged and their performances are compared against a simple historical weekly mean. Numerical results show that A-GBR performs better on average by 32% for 24-hour horizon and 39% for 12-hour horizon, whereas ExTR outdoes the other models on average by 10% for 1-hour horizon.

Sundararajan, Aditya↗
DOE OSTI2019

Learning from Dynamic User Interaction Graphs to Forecast Diverse Social Behavior

Most of the existing graph analytics for understanding social be- havior focuses on learning from static rather than dynamic graphs using hand-crafted network features or recently emerged graph embeddings learned independently from a downstream predictive task, and solving predictive (e.g., link prediction) rather than fore- casting tasks directly. To address these limitations, we propose (1) a novel task – forecasting user interactions over dynamic social graphs, and (2) a novel deep learning, multi-task, node-aware at- tention model that focuses on forecasting social interactions, going beyond recently emerged approaches for learning dynamic graph embeddings. Our model relies on graph convolutions and recurrent layers to forecast future social behavior and interaction patterns in dynamic social graphs. We evaluate our model on the ability to forecast the number of retweets and mentions of a specific news source on Twitter (focusing on deceptive and credible news sources) with R2 of 0.79 for retweets and 0.81 for mentions. An additional evaluation includes model forecasts of user-repository interactions on GitHub and comments to a specific video on YouTube with a mean absolute error close to 2% and R2 exceeding 0.69. Our results demonstrate that learning from connectivity information over time in combination with node embeddings yields better forecasting results than when we incorporate the state-of-the-art graph em- beddings e.g., Node2Vec and DeepWalk into our model. Finally, we perform in-depth analyses to examine factors that influence model performance across tasks and different graph types e.g., the influence of training and forecasting windows as well as graph topological properties.

Shrestha, Prasha↗
DOE OSTI2021

Improving the prediction of DNI with physics-based representation of all-sky circumsolar radiation

Direct normal irradiance (DNI) is often interpreted differently in ground measurements and forecasts using numerical weather prediction (NWP) models, leading to substantial bias in DNI forecasting especially under cloudy-sky conditions. To mitigate the bias, we use the Fast All-sky Radiation Model for Solar applications with DNI (FARMS-DNI) to conduct physics-based simulations of solar radiation in the circumsolar region. The DNI is quantified using the sum of the radiation along the sun direction and the scattered radiation within the circumsolar region. FARMS-DNI is coupled with the Weather Research and Forecasting model with solar extensions (WRF-Solar) to predict day-ahead DNI in 9-km pixels over an extended domain covering the contiguous United States. The DNI forecasts for the entire year of 2018 are validated using surface-based observations at the Atmospheric Radiation Measurement (ARM) and Surface Radiation Budget Network (SURFRAD) sites and a satellite-based solar radiation product developed by the procedure of the National Solar Radiation Data Base (NSRDB). The results show that the WRF-Solar coupled with the conventional Beer-Bouguer-Lambert law underestimes the overall DNI forecasts by 30%-60%, which is a joint effect of the uncertainties in computing the circumsolar radiation and forecasting cloud properties. This bias is significantly reduced by using FARMS-DNI.

14 SOLAR ENERGY↗