Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “FORECASTING”

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 361 records · Page 20
All sourcesNASA NTRSDOE OSTI
NASA NTRS2017

Assimilation of AERONET and MODIS AOT Observations Using Variational and Ensemble Data Assimilation Methods and Its Impact on Aerosol Forecasting Skill

Data assimilation of Aerosol Robotic Network (AERONET) and Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical thickness (AOT) for aerosol forecasting was tested within the Navy Aerosol Analysis Prediction System (NAAPS) framework, using variational and ensemble data assimilation methods. Navy aerosol forecasting currently makes use of a deterministic NAAPS simulation coupled to Navy Variational Data Assimilation System for aerosol optical depth, a two-dimensional variational data assimilation system, for MODIS AOT assimilation. An ensemble version of NAAPS (ENAAPS) coupled to an ensemble adjustment Kalman filter (EAKF) from the Data Assimilation Research Testbed was recently developed, allowing for a range of data assimilation and forecasting experiments to be run with deterministic NAAPS and ENAAPS. The main findings are that the EAKF, with its flow-dependent error covariances, makes better use of sparse observations such as AERONET AOT. Assimilating individual AERONET observations in the two-dimensional variational system can increase the analysis errors when observations are located in high AOT gradient regions. By including AERONET with MODIS AOT assimilation, the magnitudes of peak aerosol events (AOT> 1) were better captured with improved temporal variability, especially in India and Asia where aerosol prediction is a challenge. Assimilating AERONET AOT with MODIS had little impact on the 24 h forecast skill compared to MODIS assimilation only, but differences were found downwind of AERONET sites. The 24 h forecast skill was approximately the same for forecasts initialized with analyses from AERONET AOT assimilation alone compared to MODIS assimilation, particularly in regions where the AERONET network is dense; including the United States and Europe, indicating that AERONET could serve as a backup observation network for over-land synoptic-scale aerosol events.

Rubin, Juli I.↗
NASA NTRS2018

Evaluation of Near-Surface Air Temperature from Reanalysis over the United States and Ukraine: Application to Winter Wheat Yield Forecasting

In this work we evaluate the near-surface air temperature datasets from the ERA-Interim, JRA55, MERRA2, NCEP1, and NCEP2 reanalysis projects. Reanalysis data were first compared to observations from weather stations located on wheat areas of the United States and Ukraine, and then evaluated in the context of a winter wheat yield forecast model. Results from the comparison with weather station data showed that all datasets performed well (r2>0.95) and that more modern reanalysis such as ERAI had lower errors (RMSD ~ 0.9) than the older, lower resolution datasets like NCEP1 (RMSD ~ 2.4). We also analyze the impact of using surface air temperature data from different reanalysis products on the estimations made by a winter wheat yield forecast model. The forecast model uses information of the accumulated Growing Degree Day (GDD) during the growing season to estimate the peak NDVI signal. When the temperature data from the different reanalysis projects were used in the yield model to compute the accumulated GDD and forecast the winter wheat yield, the results showed smaller variations between obtained values, with differences in yield forecast error of around 2% in the most extreme case. These results suggest that the impact of temperature discrepancies between datasets in the yield forecast model get diminished as the values are accumulated through the growing season.

GSOD↗
NASA NTRS2020

Impact of Satellite Sea Surface Salinity Observations on ENSO Predictions from the NASA/GMAO Seasonal Forecast System

We assess the impact of satellite sea surface salinity (SSS) observations on dynamical ENSO forecasts. Assimilation of SSS improves the mixed layer depth (MLD) and modulates the Kelvin waves associated with ENSO. In column 2, the initialization differences between experiments that assimilate SSS minus those withholding SSS assimilation are presented. Column 3 shows examples of forecasts generated for the different phases of ENSO assimilating the different satellite SSS. In general, for all phases of ENSO, SSS assimilation improves forecasts. The far right column compares ensemble means for assimilation of individual and combined SMOS, Aquarius, and SMAP SSS forecasts. Finally, the latest forecasts are presented comparing assimilation versus no- assimilation of satellite SSS for single forecasts over the last year.

Hackert, E.↗
NASA NTRS2020

Designing an Optimal Ensemble Strategy for GMAO S2S Forecast System

GMAO Sub/Seasonal prediction system (S2S) is being readied for a major upgrade to GEOS-S2S Version 3. An important factor in successful extended range forecast is the definition of an ensemble For initialization of the ensemble we propose a combination of lagged and burst initial conditions. We plan to run a relatively large ensemble of 40 members for sub-seasonal forecast (up to 3 months), at which point we sub-sample the ensemble, and continue the forecast with 10 members (up to 12 months). Here we present the results of the extensive testing of various ways to generate the perturbations to the initial conditions and the validation of the stratified sampling strategy we chose.To generate perturbations for the burst ensemble members we used scaled differences of pairs of analysis states separated by 1-10 days, randomly chosen from a corresponding season. We considered perturbing separately only the atmospheric fields or only the ocean or both of the forecast initial conditions. Considering varying separation times between the analysis states, we were able to produce perturbations sampling various modes of variability. Focusing on the ENSO SST indices, we found that all types of perturbations are important for the ensemble spread.Our ensemble size for sub-seasonal forecasts was determined as to maximize the skill of predicting some of the leading modes of boreal winter atmospheric modes, NAO, PNA and AO. It is not feasible to run equally large ensemble for seasonal forecasts. Using a stratified sampling procedure we can identify the emerging directions of error growth. By comparing the stratified ensemble with randomly sampled ensemble of the same size, we were able to show that the former better estimates the mean of the original large ensemble.

Borovikov, Anna↗
NASA NTRSDate not supplied

The Added Value of SMAP Soil Moisture in Crop Yield Forecasting Over Argentina

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.

Nazmus Shams Sazib↗
NASA NTRS2020

Windows of Opportunity for Skillful Forecasts Subseasonal to Seasonal and Beyond

There is high demand and a growing expectation for predictions of environmental conditions that go beyond 0–14-day weather forecasts with outlooks extending to one or more seasons and beyond. This is driven by the needs of the energy, water management, and agriculture sectors, to name a few. There is an increasing realization that, unlike weather forecasts, prediction skill on longer time scales can leverage specific climate phenomena or conditions for a predictable signal above the weather noise. Currently, it is understood that these conditions are intermittent in time and have spatially heterogeneous impacts on skill, hence providing strategic windows of opportunity for skillful forecasts. Research points to such windows of opportunity, including El Niño or La Niña events, active periods of the Madden–Julian oscillation, disruptions of the stratospheric polar vortex, when certain large-scale atmospheric regimes are in place, or when persistent anomalies occur in the ocean or land surface. Gains could be obtained by increasingly developing prediction tools and metrics that strategically target these specific windows of opportunity. Across the globe, reevaluating forecasts in this manner could find value in forecasts previously discarded as not skillful. Users’ expectations for prediction skill could be more adequately met, as they are better aware of when and where to expect skill and if the prediction is actionable. Given that there is still untapped potential, in terms of process understanding and prediction methodologies, it is safe to expect that in the future forecast opportunities will expand. Process research and the development of innovative methodologies will aid such progress.

Annarita Mariotti↗
NASA NTRS2021

Contribution of Meteorological Downscaling to Skill and Precision of Seasonal Drought Forecasts

Research in meteorological prediction on sub-seasonal to seasonal (S2S) timescales has seen growth in recent years. Concurrent with this, demand for seasonal drought forecasting has risen. While there is obvious synergy between these fields, S2S meteorological forecasting has typically focused on low resolution global models, while the development of drought can be sensitive to the local expression of weather anomalies and their interaction with local surface properties and processes. This suggests that downscaling might play an important role in the application of meteorological S2S forecasts to skillful forecasting of drought. Here, we apply the Generalized Analog Regression Downscaling (GARD) algorithm to downscale meteorological hindcasts from the NASA Goddard Earth Observing System (GEOS) global S2S forecast system. Downscaled meteorological fields are then applied to drive offline simulations with the Catchment Land Surface Model (CLSM) to forecast United States Drought Monitor (USDM) style drought indicators derived from simulated surface hydrology variables. We compare the representation of drought in these downscaled hindcasts to hindcasts that are not downscaled, using the North American Land Data Assimilation System Phase 2 (NLDAS-2) dataset as an observational reference. We find that downscaling using GARD improves hindcasts of temperature and temperature anomalies, but the results for precipitation are mixed and generally small. Overall, GARD downscaling led to improved hindcast skill for total drought across the Contiguous United States (CONUS), and improvements were greatest for extreme (D3) and exceptional (D4) drought categories.

Ryan A Zamora↗
NASA NTRSDate not supplied

Effect of Rain-Adjusted Satellite Sea Surface Salinity on ENSO Predictions from the GMAO S2S Forecast System

ENSO has a significant impact on climate variability throughout the world and so has been the key focus for improving coupled ocean-atmosphere forecasts. Assimilation of altimetry and in situ data lead to improved short-term forecasts of the coupled system. However, few studies have focused on improving the near-surface density/mixing through assimilation of satellite sea surface salinity (SSS). For expediency, most projects that do assimilate SSS do so as if these data were observed at the top model layer instead of at the surface. In rainy regions where buoyant water sits as a fresh lens at the surface, this assumption is likely invalid. Therefore, we adjust SSS so that it more accurately represents the salinity at the first model layer. The Rain Impact Model (RIM) uses a simple diffusion model to determine the near surface salinity gradient (i.e., 1 cm to 5 m). Satellite SSS data are modified using this near-surface salinity gradient, so the salinity values are now valid at the first model layer (we call this SSS@5m). We assess the impact of satellite SSS observations for near-surface dynamics within ocean reanalyses and how these impact dynamical ENSO forecasts using the NASA GMAO S2S coupled system. For all reanalysis experiments, all available along-track sea level and in situ observations are assimilated for 2011 to 2020 using the LETKF scheme (Penny et al., 2013). One reanalysis assimilates Aquarius /SMAP SSS as before. An additional reanalysis is performed assimilating the SSS@5m data. Validation statistics are compared for experiments that assimilate SSS (sub-optimally as before) versus the SSS@5m. We also compare results of coupled forecasts that are initialized from these reanalyses in spring. For all but the big La Niña in 2017, all NINO3.4 forecasts were improved by using the RIM. We will show that improved SSS estimates upgrades density and near-surface mixing leading to more accurate coupled air/sea interaction and better forecasts.

ENSO↗
NASA NTRSDate not supplied

Exploring Coupled Land-Atmosphere Carbon Dynamics and Seasonal Carbon Forecast Skill

Terrestrial biosphere is an integral part of the climate system and plays a vital role in controlling the dynamics of the Earth’s carbon cycle. In this talk, I will present recent efforts to explore coupled land-atmosphere carbon dynamics and seasonal carbon cycle predictability. First, by imposing an idealized spring drought in coupled land‐atmosphere ensemble simulations, we investigated the impact of a regional drought on land and atmospheric carbon. Through the drought-induced impact on remote meteorology, the capability of the land’s carbon uptake is altered in both the drought area and the adjacent areas. The induced anomalies of the atmospheric CO2 extend to an area up to three times of that of the imposed drought, which suggests that atmospheric transport needs to be considered in the interpretation of the drought impact on carbon anomalies. Increase in column‐averaged monthly CO2 is at the edge of the uncertainty from single soundings of current greenhouse gas observing satellites. Secondly, the seasonal carbon forecast skill was explored using NASA’s subseasonal-to-seasonal (S2S) ensemble forecast and terrestrial biosphere model. The result demonstrates an ability to accurately predict spring-summer carbon uptake in the Northern Hemisphere mid and high latitude land at multi-month leads. The prediction skill appears to be achieved by accurate forecasts of snow removal timing, which provides a latent predictability to the forecast system, and by proper initialization of carbon and nitrogen states. Additionally, I will briefly discuss other ongoing research to improve hydrometeorological prediction by integrating biogeochemical processes, to evaluate forecast skill of wildfire carbon, and to apply subseasonal forecasts to water resources management.

Eunjee Lee↗
NASA NTRS2022

Improving Multiday Solar Wind Speed Forecasts

We analyze the residual errors for the Wang-Sheeley-Arge (WSA) solar wind speed forecasts as a function of the photospheric magnetic field expansion factor (fp) and the minimum separation angle (d) in the photosphere between the footpoints of open field lines and the nearest coronal hole boundary. We find the map of residual speed errors are systematic when examined as a function of fp and d. We use these residual error maps to apply corrections to the model speeds. We test this correction approach using 3-day lead time speed forecasts for an entire year of observations and model results. Our methods can readily be applied to develop corrections for the remaining WSA forecast lead times which range from 1 to 7 days in 1-day increments. Since the solar wind density, temperature, and the interplanetary magnetic field strength all correlate well with the solar wind speed, the improved accuracy of solar wind speed forecasts enables the production of multiday forecasts of the solar wind density, temperature, pressure, and interplanetary field strength, and geophysical indices. These additional parameters would expand the usefulness of Air Force Data Assimilative Photospheric Flux Transport-WSA forecasts for space weather clients.

H A Elliot↗
NASA NTRSDate not supplied

Inferred Sea Level Prediction in the NASA GMAO Seasonal Forecasting System

Reliable predictions of sea level anomalies on seasonal timescales with lead times of 1 to 9 months may have relevance to stakeholders – for example, in the advance deployment of resources for coastal flood mitigation. Routine prediction and analysis may also highlight physical processes associated with sea level change and modeling capabilities on seasonal and other timescales. These forecasts may represent interannual changes in the seasonal slope of the ocean surface, teleconnection effects such as the El Niño/Southern Oscillation phenomenon, and variations in seasonal hydrology including precipitation and coastal runoff. Coupled atmosphere/ocean models are routinely used in the seasonal prediction of temperature anomalies, precipitation anomalies, sea ice cover, and climate indices such as the Niño3.4 predictions under the North American Multi-Model Ensemble (NMME) protocol. Within the limits of their configuration, these complex Earth-system models have a potential for depicting regional changes in oceanic column properties, including the sea surface height. Seasonal prediction models generally have no representation of long-term mass contributions from melting land ice, or changes in vertical land motion; their output may be more specifically characterized as predictions of the ocean dynamic sea level. In practice however, the sea surface height prognostic variable is substantially compromised by the forecast model response to initial conditions. Imbalances between the initial, observed hydrologic cycle and the forecast model state produce abrupt adjustments in the model sea surface height. As a result, most seasonal prediction systems employ a constraint on the globally-averaged sea surface height that is applied at each time step. This essentially renders the prognostic sea surface height variable as unserviceable. Several approaches have previously been used to retrieve sea level information from seasonal forecasts beyond the use of the sea surface height variable. Here, we extend a method of relating other prognostic values, including ocean circulation and climate indices, to observed sea level variations. We use the merged altimetry record of the NASA MEaSUREs Gridded Sea Surface Height Anomalies data set and monthly revised local reference gauge observations from the National Oceanography Centre Permanent Service for Mean Sea Level (PSMSL) to evaluate derived prognostic variables from the NASA Global Modeling and Assimilation Office subseasonal-to-seasonal system version 2.1 (GMAO S2S v2.1). We focus on results for the midlatitudes with particular emphasis on US gauge locations. As shown in previous studies, prognostic ENSO-related indices in boreal winter are well correlated with gauge observations for the US west coast, but also for other locations in the southeastern US. Other forecast climate indices such as the North Atlantic Oscillation have relations to sea level that are limited both seasonally and spatially. As expected, surface atmospheric pressure (e.g., inverse barometer effect) is found to be particularly well correlated with observed sea level. We provide a characterization of forecast skill for seasonal sea level with this method.

Richard I Cullather↗
NASA NTRSDate not supplied

Challenges in Observing, Modeling, and Forecasting the June 2023 Smoke Event Over the Northeast United States

In the first week of May 2023, boreal Canada began an early start to the biomass burning season when fires erupted across Alberta and western Saskatchewan. Smoke entered the troposphere and spread across the United States as biomass burning emissions in Canada quadrupled the prior maximum since the onset of the MODIS satellite record. Adding to the already existing smoke, wildfires ignited in Quebec on June 2, 2023, producing a heavy plume of aerosol over the northeastern part of the continent. Meteorological factors, including an area of low pressure situated, nearly stationary, over Maine, transported the smoke into a densely populated corridor containing the cities of Washington, DC, Baltimore, Philadelphia, and New York City. Air quality alerts became widespread as near-surface levels of fine particulate matter (PM2.5) exceeded harmful levels due the low altitude of the smoke plume. Forecasting the smoke transport and impacts of such events requires a complex combination of observed initial conditions for meteorology and aerosols, knowledge of the emissions from wildfires, and a state-of-the-art Earth System model that couples these components together. Using a case-study perspective with the Goddard Earth Observing System (GEOS), challenges associated with near real time forecasting of the smoke plume are presented. Analyzed and forecasted aerosol optical depth will be evaluated using available satellite and AERONET observations, while surface aerosol will be assessed relative to observations of PM2.5. Lacking real time estimates of biomass burning emissions for inputs to our forecast model, particular attention is given to the use of day-old biomass burning emissions throughout the forecast period covering the first ten days in June 2023, and the resulting reduction in forecast skill will be quantified.

Allison Collow↗
NASA NTRSDate not supplied

Challenges in Observing, Modeling, and Forecasting the June 2023 Smoke Event over the Northeast United States

In the first week of May 2023, boreal Canada began an early start to the biomass burning season when fires erupted across Alberta and western Saskatchewan. Smoke entered the troposphere and spread across the United States as biomass burning emissions in Canada quadrupled the prior maximum since the onset of the MODIS satellite record. Adding to the already existing smoke, wildfires ignited in Quebec on June 2, 2023, producing a heavy plume of aerosol over the northeastern part of the continent. Meteorological factors, including an area of low pressure situated, nearly stationary, over Maine, transported the smoke into a densely populated corridor containing the cities of Washington, DC, Baltimore, Philadelphia, and New York City. Air quality alerts became widespread as near-surface levels of fine particulate matter (PM2.5) exceeded harmful levels due the low altitude of the smoke plume. Forecasting the smoke transport and impacts of such events requires a complex combination of observed initial conditions for meteorology and aerosols, knowledge of the emissions from wildfires, and a state-of-the-art Earth System model that couples these components together. Using a case-study perspective with the Goddard Earth Observing System (GEOS ), challenges associated with near real time forecasting of the smoke plume are presented. Analyzed and forecasted aerosol optical depth will be evaluated using available satellite and AERONET observations, while surface aerosol will be assessed relative to observations of PM2.5. Lacking real time estimates of biomass burning emissions for inputs to our forecast model, particular attention is given to the use of day-old biomass burning emissions throughout the forecast period covering the first ten days in June 2023, and the resulting reduction in forecast skill will be quantified.

Allison Collow↗
DOE OSTI2022

Local-thermal-gradient and large-scale-circulation impacts on turbine-height wind speed forecasting over the Columbia River Basin

Abstract. We investigate the sensitivity of turbine-height wind speed forecast to initial condition (IC) uncertainties over the Columbia River Gorge (CRG) and Columbia River Basin (CRB) for two typical weather phenomena, i.e., local-thermal-gradient-induced marine air intrusion and a cold frontal passage. Four types of turbine-height wind forecast anomalies and their associated IC uncertainties related to local thermal gradients and large-scale circulations are identified using the self-organizing map (SOM) technique. The four SOM types are categorized into two patterns, each accounting for half of the ensemble members. The first pattern corresponds to IC uncertainties that alter the wind forecast through a modulating weather system, which produces the strongest wind anomalies in the CRG and CRB. In the second pattern, the moderate uncertainties in local thermal gradient and large-scale circulation jointly contribute to wind forecast anomaly. We analyze the cross section of wind and temperature anomalies through the gorge to explore the evolution of vertical features of each SOM type. The turbine-height wind anomalies induced by large-scale IC uncertainties are more concentrated near the front. In contrast, turbine-height wind anomalies induced by the local IC thermal uncertainties are found above the surface thermal anomalies. Moreover, the wind forecast accuracy in the CRG and CRB is limited by IC uncertainties in a few specific regions, e.g., the 2 m temperature within the basin and large-scale circulation over the northeast Pacific around 140∘ W.

17 WIND ENERGY↗
DOE OSTI2023

Using High-Resolution NSRDB Data to Evaluate Cloud Mask Forecast from WRF-Solar EPS

Validating spatiotemporal distributions of cloud forecasts using numerical weather prediction (NWP) models is difficult as this requires high-quality cloud-property at significantly high spatial and temporal resolution over extended periods of time. Observations of cloud properties, such as cloud mask, cloud optical thickness, and cloud type, are vital for assessing the capability of NWP models to forecast various types of clouds. Using the National Solar Radiation Database (NSRDB), this research evaluates ensemble cloud-mask predictions from the WRF-Solar ensemble prediction system (WRF-Solar EPS). From the WRF-Solar EPS, day-ahead solar forecasts for the contiguous United States (CONUS) for 2018 are simulated. Given the NSRDB data is accessible at a resolution of 2 km, we can calculate the cloud fraction across the 9-km grid of WRF-Solar EPS. This allows us to spatially assess the cloud-mask forecasts using two methods against the high-resolution NSRDB: (a) considering all 2-km NSRDB clouds in the forecast domain (EMAll), and (b) using a minimum cloud fraction threshold of 50% to designate a pixel as cloudy (EMP50). The low-resolution cloud masks from WRF-Solar EPS are evaluated directly against the cloud-resolving scale gridded observations from NSRDB using EMAll. With EMP50, we presume that scenes with less than 50% cloud cover from the 2-km NSRDB are clear. Thus, this assessment approach allows for a fair comparison with WRF-Solar EPS resolved to a 9-km grid. A method of point-by-point verification is used to evaluate dichotomous (yes/no) cloud mask predictions against the NSRDB. For each pixel of model extent, cloud frequency and traditional metrics (e.g., probability of detection, false alarm rate, and hit rate, etc.) are computed and compared with satellite-derived data sets. Mismatched cloud frequency (MCF) is computed to measure the present capability of WRF-Solar EPS in representing various types of clouds, which are categorized using three levels of cloud top height (CTH) and cloud optical depth (COD) across entire CONUS. Preliminary results show that the WRF-Solar EPS provides MCF values ranging from 9% to 46%, 16% to 33%, and 8% to 27% for low-level, middle-level, and high-level clouds, respectively, for three CTHs. The model produces MCFs ranging from 27% to 46%, 13% to 34%, and 8% to 19% for thin, medium-thickness, and thick clouds, respectively, for three CODs. The presentation will include a detailed description of the current outcomes as well as potential future extensions. The evaluation approach established in this study is readily extensible to the evaluation of cloud predictions from different ensemble NWP models. In addition, the findings of the suggested evaluation technique aid in identifying model weaknesses and will ultimately lead to advances in WRF-Solar EPS's skill in predicting clouds and solar irradiance.

cloud mask forecast↗
DOE OSTI2021

Review of Onsite Temperature and Solar Forecasting Models to Enable Better Building Design and Operations

Advanced building controls and energy optimization for new constructions and retrofits rely on accurate weather data. Traditionally, most studies utilize airport weather information as the decision inputs. However, most buildings are in environments that are quite different than those at the airport miles away. Tree cover, adjacent buildings, and micro-climate effects caused by the larger surrounding area can all yield deviations in air temperature, humidity, solar irradiance, and wind that are large enough to influence design and operation decisions. In order to overcome this challenge, there are many prior studies on developing weather forecasting algorithms from micro-to meso-scales. Additionally, this paper reviews and complies knowledge on common weather data resources, data processing methodologies and forecasting techniques of weather information. Commonly used statistical, machine learning and physical-based models are discussed and presented as two major categories: deterministic forecasting and probabilistic forecasting. Finally, evaluation metrics for forecasting errors are listed and discussed.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2021

Active learning strategy for high fidelity short-term data-driven building energy forecasting

The quality of a data-driven model is heavily dependent on the quality of data. Data from building operation often have data bias problems, which means that the data sample is collected in a way that some members of the intended data population are less likely to be included than others. Data-driven energy forecasting models built on such data hence are biased and could lead to large forecasting errors. Active learning—an effective method to defying data bias—is rarely studied or applied in the area of data-driven building energy forecasting modeling. This paper attempts to fill this gap and explores the application of active learning in data-driven building energy forecasting. The developed strategy in this paper efficiently generate informative training data within a time budget and uses block design to passively consider weather disturbances. The developed active learning strategy is applied and evaluated in both virtual and real-building testbeds against traditional data-driven methods. Via these virtual and real-building evaluation cases, we have demonstrated that the data bias problem typically exists in building operation data is resolved by applying the developed active learning strategy. Furthermore, building energy forecasting models trained from data generated from the active learning strategy have shown improved performances in both model accuracy and model extendibility perspectives. The effectiveness of the block design module is also validated to effectively consider the impact of weather conditions on active learning design.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Explainable machine learning model for multi-step forecasting of reservoir inflow with uncertainty quantification

We propose an explainable machine learning (ML) model with uncertainty quantification (UQ) to improve multi-step reservoir inflow forecasting. Traditional ML methods have challenges in forecasting inflows multiple days ahead, and lack explainability and UQ. To address these limitations, we introduce an encoder–decoder long short-term memory (ED-LSTM) network for multi-step forecasting, employ the SHapley Additive exPlanation (SHAP) technique for understanding the influence of hydrometeorological factors on inflow prediction, and develop a novel UQ method for prediction trustworthiness. We apply these methods to forecast 7-day inflow in snow-dominant and rain-driven reservoirs. The results demonstrate the effectiveness of the ED-LSTM model, with high forecasting accuracy for short lead times. Our UQ method provides reliable uncertainty estimates, covering 90% of data with a 90% confidence level. The SHAP analysis reveals the importance of historical inflow and precipitation as influential factors. These findings and methods may support reservoir operators in optimizing water resources management decisions.

54 ENVIRONMENTAL SCIENCES↗