Results for “wind direction variability” — Page 2

NASA NTRS2018

Magnetotail Boundary Crossings at Lunar Distances: ARTEMIS Observations

We compare results from a preliminary analysis of two years of ARTEMIS magnetopause boundary crossings at lunar distances with available empirical models. We remove the effects of variable solar wind flow directions and aberration angles to study the magnetotail cross-section as a function of solar wind conditions. The average magnetopause distance from the central axis is 26 RE, but this distance ranges from 10 RE for high solar wind dynamic pressures and strong northward IMF orientations to 39 RE for low solar wind dynamic pressures and weak southward IMF orientations compared to the nominal solar wind conditions. The time-independent Howe and Binsack (1972) model describes the average location of the crossings very well. For high solar wind dynamic pressures, the Lu et al., (2011) model performs best, while for low pressures the Petrinec and Russell (1996) model gives the closest prediction. As predicted by theory and seen in past studies, the magnetotail cross-section is suggestive of prolate during intervals of strong IMF By, but oblate during intervals of strong IMF Bz. Any asymmetric variation of the tail boundary with respect to the sign of IMF By was not observed. The decreasing size of boundary with the increasing dynamic pressure was found when dynamic pressures are smaller than 2nPa. Although the scatter is larger, the tail size for pressures larger than 2nPa suggests a constant radius. The tail boundary size decreases as the strength of IMF Bz increases regardless of its polarity. However, it was also observed that an even stronger southward IMF Bz can cause larger magnetopause size in the presence of large dynamic pressures.

Magnetotail↗

NASA NTRSDate not supplied

Predicting Airport Runway Configurations for Decision-Support Using Supervised Learning

One of the most challenging tasks for air traffic controllers is runway configuration management (RCM). It deals with the optimal selection of runways to operate on (for arrivals and departures) based on traffic, surface wind speed, wind direction, other environmental variables, noise constraints, and several other airport-specific factors. It affects the efficiency of the National Airspace System (NAS) and both surface and airspace operations can benefit from better understanding future runway configurations. In this paper, we present a comprehensive implementation of predictive models for runway configuration estimation from large volumes of historical data. Specifically, operational data from two full years (2018 and 2019) is collected, analyzed, and fused together to build the data product used in this work. The data set differs from prior work in the field in terms of its scope, resolution, and variety of factors collected and considered. Meteorological data is collected from two different sources – current weather conditions from METAR (Meteorological Terminal Aviation Routine Weather Report) and forecast weather conditions from Localized Aviation MOS Program (LAMP). Operational data from the Federal Aviation Administration (FAA) Aviation System Performance Metrics (ASPM) related to scheduled and actual number of arrivals and departures, average taxi times, etc. are collected. NASA’s Sherlock Data Warehouse is used to identify critical information such as go-arounds, and other events that might impact RCM decision-making. All data is collected and aggregated over 15-minute intervals throughout the two years. This provides a resolution like the timescales that might be necessary for runway configuration management decision-making. A variety of supervised learning algorithms are tested including Support Vector Machine, Random Forest, Gradient Boosting, etc. including tuning of the model hyperparameters. The modeling process is applied and presented on two representative U.S. airports – Charlotte Douglas International Airport (KCLT) and Denver International Airport (KDEN). The two airports present different levels of complexity in terms of the total number of configurations used and provide a balanced perspective on the generalizability of the developed approach to other airports in the NAS. Initial results are promising (F1 score of 0.91 at KCLT and 0.83 at KDEN) for data in the test set. The final paper will contain a comprehensive comparison between different models and model building strategies as well as further refined results. Most important predictors for each airport will be identified along with a discussion and recommendations on adapting the framework to other scenarios.

Tejas G Puranik↗

NASA NTRSDate not supplied

Predicting Airport Runway Configurations for Decision-Support Using Supervised Learning

One of the most challenging tasks for air traffic controllers is runway configuration management (RCM). It deals with the optimal selection of runways to operate on (for arrivals and departures) based on current and forecast of traffic, surface wind speed, wind direction, other environmental variables, noise constraints, and several other airport-specific factors. In this paper, a methodology using supervised learning is developed to build a predictive model for RCM decision-support from large volumes of historical data. Data from two full years (2018 and 2019) related to current and forecast weather, demand/capacity, etc. is collected, analyzed, and fused together. A variety of supervised learning algorithms are tested for predicting runway configuration and hyperparameter tuning is carried out to select the best performing model. The validation process involves two airports of low (Charlotte Douglas International Airport, CLT) and high (Denver International Airport, DEN) complexity of configuration decision-making. The results show significant promise for the two airports with test accuracy of 93% (CLT) and 73% (DEN). The methodology is scalable and generalizable to other airports across the U.S. National Airspace System.

air traffic management↗

NASA NTRSDate not supplied

Predicting Airport Runway Configuration for Decision-Support Using Supervised Learning

One of the most challenging tasks for air traffic controllers is runway configuration management (RCM). It deals with the optimal selection of runways to operate on (for arrivals and departures) based on current and forecast of traffic, surface wind speed, wind direction, other environmental variables, noise constraints, and several other airport-specific factors. In this paper, a methodology using supervised learning is developed to build a predictive model for RCM decision-support from large volumes of historical data. Data from two full years (2018 and 2019) related to current and forecast weather, demand/capacity, etc. is collected, analyzed, and fused together. A variety of supervised learning algorithms are tested for predicting runway configuration and hyperparameter tuning is carried out to select the best performing model. The validation process involves two airports of low (Charlotte Douglas International Airport, CLT) and high (Denver International Airport, DEN) complexity of configuration decision-making. The results show significant promise for the two airports with test accuracy of 93% (CLT) and 73% (DEN). The methodology is scalable and generalizable to other airports across the U.S. National Airspace System.

air traffic management↗

DOE OSTI2025

Key Environmental and Ecological Variables of Wetland CH 4 and CO 2 Fluxes Change With Warming

Wetlands are important ecosystems for the global carbon cycle, impacting regional and global methane (CH 4 ) and carbon dioxide (CO 2 ) budgets. This study examines how environmental and ecological variables impact wetland CH 4 flux and net ecosystem exchange of CO 2 (NEE) across 17 sites globally. We also quantified the importance of variables for each wetland type and site at monthly scale under normal and warm temperatures using dominance analysis. We identified soil and air temperature (TS, TA, respectively) as key variables influencing wetland CH4, and latent heat (LE) and shortwave radiation (SW) for NEE under normal and warm conditions. However, the importance of some variables shifted with warming. For predicting the variability of wetland CH4 flux under warming, gross primary productivity (GPP) and LE, replacing wind direction (WD), were dominant variables for tropical swamps, while NEE was important for high-latitude fens and bogs under warm temperatures. For wetland NEE, the role of TA and TS decreased across all wetland types with warming, while vapor pressure deficit (VPD) became more important for mid and high-latitude wetlands. Our results reveal the complex responses of wetland carbon flux to environmental and ecological variables with warming and provide new insights into improving wetland models by incorporating additional variables and accounting for the changing roles of variables in carbon flux under warming.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2019

Significant Spatial Variability in Radar-Derived West Antarctic Accumulation Linked to Surface Winds and Topography

Across the Antarctic Ice Sheet, accumulation heavily influences firn compaction and surface height changes. Therefore, accumulation varies over short distances (<25 km), complicating the derivation of ice sheet mass changes from altimetry and reducing how accurately field measurements can be spatially extrapolated. However, current atmospheric reanalyses have grid spacings (>25 km) that are too coarse to resolve this variability. To address this limitation, we construct a fine-scale accumulation product from airborne snow radar observations by superimposing along-track fluctuations in accumulation onto an atmospheric reanalysis product. Our resulting airborne product reflects large-scale (>25 km) orographic precipitation patterns while providing robust and unprecedented insight into Antarctic accumulation variability on subgrid scales. On these smaller scales, we find significant, regionally dependent accumulation variability (𝜎(sub relative) > 40%). This variability in accumulation is correlated with variability in topographic surface slope in the wind direction (p < 0.01), confirming that subgrid-scale accumulation variability is driven by snow redistribution by wind.

Dattler, Marissa E.↗

DOE OSTI2026

Meteorological and Soil Data from Ecohydrology Sensor Towers at Pump House and Snodgrass Mountain in East River Watershed, Colorado, 2019-2025

This data package includes hourly meteorological and soil sensor data at eight ecohydrology monitoring sites in East River Watershed, Colorado as part of the Watershed Function Scientific Focus Area (WFSFA) research led by Lawrence Berkeley National Lab (LBNL). Four field sites were located on the hillslope of East River (ER) near Pump House (PH) at Mount Crested Butte (ER-PHS1 to 4), and the other four are in the Snodgrass Mountain (SG) area (SG-EHS5 to 8). In terms of vegetation cover, three sites are in montane grasslands (ER-PHS1, ER-PHS2, and SG-EHS5), three are below evergreen conifer canopy (ER-PHS3, SG-EHS6, and SG-EHS7), and two are below deciduous aspen canopy (ER-PHS4 and SG-EHS8). The monitoring period began in October 2019 at the East River sites, in October 2020 at SG-EHS5 and SG-EHS6, and in October 2021 at SG-EHS7 and SG-EHS8. In September 2024, all four East River sites were fully retired. The four Snodgrass Mountain sites remain active. Each site is equipped with a comprehensive suite of meteorological sensors on a tripod and soil sensors that measure weather, energy fluxes, and soil variables. This data package includes measurements from ten different types of sensors and up to thirteen individual sensors per site, including (1) a weather station (measurement height ranges from 2.8~3.8 meters (m) above ground), (2) a quantum sensor for photosynthetic active radiation (PAR) (2.4~3.3m), (3) a net radiometer (1.7~2.1m), (4) an infrared radiometer (1.6~2.2m), (5) a sonic distance sensor (1.5~1.9m), (6) a soil carbon dioxide (CO2) flux chamber (0m), (7) a soil heat flux plate (-0.05m below ground), (8) a soil oxygen sensor (-0.3m), (9) a soil water potential sensor (-0.3m), and (10) soil water content sensors at 3~4 depths (-1.15 ~ -0.1m). A total of twenty-three variables is reported in this data package, including (1) atmospheric variables: air temperature (TA), atmospheric pressure (PA), vapor pressure (VP), and vapor pressure deficit (VPD), (2) precipitation variables: rain precipitation (P) and snow depth (D_SNOW), (3) energy fluxes variables: four-component net radiation (NETRAD) (shortwave/longwave incoming/outgoing radiation, SW_IN, SW_OUT, LW_IN, LW_OUT), photosynthetic photon flux density (PPFD), and soil heat flux (G), (4) soil variables: soil water content (SWC), soil water potential (SWP), soil temperature (TS), soil bulk electrical conductivity (COND_SOIL), and soil gaseous oxygen concentration (O2_SOIL), (5) wind variables: two-dimensional wind speed (WS), gust speed (WS_MAX), and wind direction (WD), and (6) surface variables: surface infrared temperature (T_CANOPY) and soil CO2 flux (CO2_SOIL). Please see the Methods section for data processing and QA/QC steps taken to generate the hourly datasets. The following files are included in this data package (notes on version: v{x}-{y}, where x is the metadata version, and y is the data version, when applicable): (1) “metadata_site_v{x}-{y}.csv” - a site metadata file that summarizes location information of all sites, including site ID, description, coordinates, timeframe, elevation, and vegetation cover, (2) “metadata_instrument_v{x}-{y}.csv” - an instrument metadata file that summarizes sensor information of all sites, including sensor manufacturer and model, measurement height, and sampling and averaging interval of all variables, (3) "data_{SITE_ID}_v{x}-{y}.csv" - eight data files that contain hourly data of each site indicated by {SITE_ID} in the filename, (4) “/figure/data_{SITE_ID}_v{x}-{y}.png" - eight figures that help visualize data of each site indicated by {SITE_ID} in the filename, (5) “/photo/*” - photos of each site indicated by {SITE_ID} in the filename, and (6) four file level metadata (flmd.csv) and data dictionary (*_dd.csv) files that summarize file, header, column, and variable information of all files. Notes: (1) Measurement height: Each variable name is followed by conventional positional qualifiers “H_V_R”, where H indicates the relative horizontal positions of that specific variable, V the vertical positions, and R the replicates. In this data package, only the vertical qualifier V varies, and V increases from the highest vertical position (V=1) to the lowest. Variables with the same qualifier are not necessarily measured by the same sensor, and the same variable with the same qualifier across different sites are not necessarily measured at the same height. Please refer to “metadata_instrument.csv” for the sensor information and measurement heights, and whether a variable is measured below the canopy. (2) Variable availability: Snow depth is not available at ER-PHS3 and SG-EHS7. SWC, soil temperature, and soil bulk EC at the deepest depth (<-1m) are not available at SG-EHS6 and SG-EHS7. The missing value code for numeric variables is -9999, except for SWP. For SWP, the missing value code is +9999, because SWP values are negative. (3) Sampling frequency: Please refer to “metadata_instrument.csv” for the increase of sampling frequency of some variables from 30-min to 1-min at ER-PHS1 to 4 in July 2020. (4) Sensors: While the methods of each sensor are not detailed, all sensors are commercially available, and their methods can be found in their manuals. Please refer to “metadata_instrument.csv” for the sensor manufacturer and model information. This work was supported by the Watershed Function Science Focus Area at Lawrence Berkeley National Laboratory funded by the US Department of Energy, Office of Science, Biological and Environmental Research under Contract No. DE-AC02-05CH11231.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2021

Probabilistic Day-Ahead Forecasting Using an Analog Ensemble Approach for Wind Farm Grid Services

Wind resource assessment and wind power forecasting are used in research and industry to anticipate future power output at scales ranging from individual wind turbines to entire wind farms. Probabilistic day-ahead wind forecasting is useful for anticipating how a wind farm could potentially participate in the day-ahead market by providing upper and lower bounds for expected power generation, thus informing grid operators of its uncertainty. Understanding this uncertainty is part of a larger project focused on building a platform that combines efforts in weather forecasting, aerodynamic and economic modeling to create maximum value of a wind plant to better provide services to the grid. This effort is also known as the Atmosphere to Electrons to Grid (A2E2G) project. One method for producing a probabilistic forecast is through the analog ensemble approach (Delle Monache et al., 2011). This method leverages historical forecasts and their corresponding observations as a training data set from which future forecasts can be made. For some future forecast, the most similar historical forecasts (analogs) are identified on a regular time basis such as once per a 3-hour window. The most similar analogs, based on a metric such as root mean square error (RMSE), are recorded and their corresponding verifying observations are used as an ensemble member for this future forecast. Prior work in this area demonstrates improvements over raw Numerical Weather Prediction (NWP) forecasts and shows skill similar to techniques such as logistic regression and machine learning (Delle Monache et al., 2013; Alessandrini et al., 2015). Here, we take the High-Resolution Rapid Refresh model (HRRR) day-ahead forecast (0-36 hours) to create a probabilistic day-ahead forecast using an analog ensemble approach. The HRRR has an hourly temporal resolution, with a spatial resolution of 3 km. The 12 UTC HRRR model run is downloaded every day for one year from August 2019 - July 2020, with the first 11 months serving as a bank of analogs from which the forecasting algorithm can create a probabilistic forecast. Once downloaded, the original HRRR forecast is temporally interpolated to 5-minutes, aligning with both the temporal resolution of the observations as well as the timescale relevant for day-ahead power forecasts. The forecast is validated at the M2 tower at the Flatirons Campus of the National Renewable Energy Laboratory (NREL) at a typical wind turbine height of 80 m. Variables such as wind speed, wind direction, and turbulence intensity are incorporated into the probabilistic forecast model and weighted according to their relative importance to the forecast. Based on metrics such as mean bias error (MBE), mean absolute error (MAE), and root mean square error, the analog ensemble forecast outperforms the raw HRRR forecast during the testing period of July 2020. Figure 1 illustrates an example day-ahead forecast compared against the verifying observations. The general variability and ramps are captured throughout the day, with potential to further improve the analog ensemble model through machine learning techniques.

numerical weather prediction↗

NASA NTRS2004

Tower Mesonetwork Climatology and Interactive Display Tool

Forecasters at the 45th Weather Squadron and Spaceflight Meteorology Group use data from the tower network over the Kennedy Space Center (KSC) and Cape Canaveral Air Force Station (CCAFS) to evaluate Launch Commit Criteria, and issue and verify forecasts for ground operations. Systematic biases in these parameters could adversely affect an analysis, forecast, or verification. Also, substantial geographical variations in temperature and wind speed can occur under specific wind directions. To address these concerns, the Applied Meteorology Unit (AMU) developed a climatology of temperatures and winds from the tower network, and identified the geographical variation and significant tower biases. The mesoclimate is largely driven by the complex land-water interfaces across KSC/CCAFS. Towers with close proximity to water typically had much warmer nocturnal temperatures and higher wind speeds throughout the year. The strongest nocturnal wind speeds occurred from October to March whereas the strongest mean daytime wind speeds occurred from February to May. These results of this project can be viewed by forecasters through an interactive graphical user interface developed by the AMU. The web-based interface includes graphical and map displays of mean, standard deviation, bias, and data availability for any combination of towers, variables, months, hours, and wind directions.

Case, Jonathan L.↗

NASA NTRS2022

Low-Level Jets and the Convergence of Mars Data Assimilation Algorithms

Data assimilation is an increasingly popular technique in Mars atmospheric science, but its effect on the mean states of the underlying atmosphere models has not been thoroughly examined. The robustness of results to the choice of model and assimilation algorithm also warrants further study. We investigate these issues using two Mars general circulation models (MGCMs), with particular emphasis on zonal wind and temperature fields. When temperature retrievals from the Mars Global Surveyor Thermal Emission Spectrometer (TES) are assimilated into the U.K.-Laboratoire de M ́et ́eorologie Dynamique (UK-LMD) MGCM to create the Mars Analysis Correction Data Assimilation (MACDA) reanalysis, low-level zonal jets in the winter northern hemisphere shift equatorward and weaken relative to a free-running control simulation from the same MGCM. The Ensemble Mars Atmosphere Reanalysis System (EMARS) reanalysis, which is also based on TES temperature retrievals, also shows jet weakening (but less if any shifting) relative to a control simulation performed with the underlying Geophysical Fluid Dynamics Laboratory (GFDL) MGCM. Examining higher levels of the atmosphere, monthly mean three-dimensional temperature and zonal wind fields are in generally better agreement between the two reanalyses than between the two control simulations. In conjunction with information about the MGCMs’ physical parametrizations, intercomparisons between the various reanalyses and control simulations suggest that overall the EMARS control run is plausibly less biased (relative to the true state of the Martian atmosphere) than the MACDA control run. Implications for future observational studies are discussed. Plain Language Summary An increasingly popular way to study Martian weather and climate is to combine atmospheric temperature observations with a computer model (specifically, a global climate model). The process of combining model and observations is called “data assimilation”, and the resulting merged data set is called a “reanalysis”. One advantage of reanalyses is that they include variables (such as wind) that are not directly observed. For scientific and practical applications we want these variables to be reasonably accurate— however, it is not clear how well data assimilation algorithms compute them. Our study investigates this issue using two Mars reanalyses and two model simulations that do not assimilate temperature data. We focus on slowly-varying atmospheric phenomena (timescales from 10 Mars days to a season). Assimilating temperature data into two different global climate models changes the strength and/or spatial pattern of east-west winds at low altitudes. Furthermore, monthly mean three-dimensional temperature and east-west wind fields agree better between reanalyses than between non-assimilating model simulations. This suggests that the data assimilation process is basically successful. One non-assimilating model simulation has less realistic representations of atmospheric physical processes than the other—we argue that this plausibly gives it larger biases relative to the true state of the atmosphere.

Todd A Mooring↗

DOE OSTI2026

Short‐Term Hourly Weather Forecasting Using PredRNN With Image Preprocessing

Global weather forecast models are vital tools with numerous applications, including public safety, agriculture, and transportation. Recent advancements in artificial intelligence (AI) and deep learning (DL) have shown the potential to enhance weather forecasting accuracy and speed. In this study, we developed a short-term hourly weather forecast framework with a wavelet transform function for data preprocessing and a spatiotemporal DL model, PredRNN, for predicting five surface atmospheric variables, including wind speed and direction, mean sea level pressure (MSLP), temperature, and precipitation. The framework demonstrated promising results. It produces global forecasts at 0.25° (∼25 km) with a 1-day lead time RMSE of 1.8 m/s for wind components, 180 Pa for MSLP, and 1.8 K for temperature. Although our model does not surpass state-of-the-art AI weather forecast models across all metrics, it outperforms these models in precipitation forecasting and wind prediction at short lead times and achieves comparable accuracy for MSLP. Its native hourly forecasting capability, together with training on widely accessible GPU hardware, contributes meaningfully to the advancement of accessible DL weather forecasting methods. Our work highlights the importance of integrating temporal components and data transformation techniques to improve the predictability and accuracy of weather forecasts.

Tran, Hoang [Pacific Northwest National Laboratory↗

DOE OSTI2024

Sounding System / Raw Data

This dataset contains twice or four times daily radiosonde observations of the atmosphere from the quarterly LTER ship transects. Variables include atmospheric pressure, temperature, relative humidity, wind speed, wind direction, north wind speed, and east wind speed. Note that variables are a function of height.

17 WIND ENERGY↗

DOE OSTI2024

Sounding System / Raw Data

This dataset contains twice or four times daily radiosonde observations of the atmosphere at the Woods Hole weather station. Variables include atmospheric pressure, temperature, relative humidity, wind speed, wind direction, north wind speed, and east wind speed. Note that variables are a function of height.

17 WIND ENERGY↗

NASA NTRS1980

Wind wheel electric power generator

Wind wheel electric power generator apparatus includes a housing rotatably mounted upon a vertical support column. Primary and auxiliary funnel-type, venturi ducts are fixed onto the housing for capturing wind currents and conducting to a bladed wheel adapted to be operatively connected with the generator apparatus. Additional air flows are also conducted onto the bladed wheel; all of the air flows positively effecting rotation of the wheel in a cumulative manner. The auxiliary ducts are disposed at an acute angle with respect to the longitudinal axis of the housing, and this feature, together with the rotatability of the housing and the ducts, permits capture of wind currents within a variable directional range.

Kaufman, J. W.↗

NASA NTRS1986

Prediction of the far field noise from wind energy farms

The basic physical factors involved in making predictions of wind turbine noise and an approach which allows for differences in the machines, the wind energy farm configurations and propagation conditions are reviewed. Example calculations to illustrate the sensitivity of the radiated noise to such variables as machine size, spacing and numbers, and such atmosphere variables as absorption and wind direction are presented. It is found that calculated far field distances to particular sound level contours are greater for lower values of atmospheric absorption, for a larger total number of machines, for additional rows of machines and for more powerful machines. At short and intermediate distances, higher sound pressure levels are calculated for closer machine spacings, for more powerful machines, for longer row lengths and for closer row spacings.

Shepherd, K. P.↗

NASA NTRS2000

A Flight Training Simulator for Instructing the Helicopter Autorotation Maneuver (Enhanced Version)

Autorotation is a maneuver that permits a safe helicopter landing when the engine loses power. A catastrophe may occur if the pilot's control inputs are incorrect, insufficient, excessive, or poorly timed. Due to the danger involved, full-touchdown autorotations are very rarely practiced. Because in-flight autorotation training is risky, time-consuming, and expensive, the objective of the project was to develop the first helicopter flight simulator expressly designed to train students in this critical maneuver. A central feature of the project was the inclusion of an enhanced version of the Pilot-Rotorcraft Intelligent Symbology Management Simulator (PRISMS), a virtual-reality system developed by Anacapa Sciences and Thought Wave. A task analysis was performed to identify the procedural steps in the autorotation, to inventory the information needed to support student task performance, to identify typical errors, and to structure the simulator's practice environment. The system provides immediate knowledge of results, extensive practice of perceptual-motor skills, part-task training, and augmented cueing in a realistic cockpit environment. Additional work, described in this report, extended the capabilities of the simulator in three areas: 1. Incorporation of visual training aids to assist the student in learning the proper appearance of the visual scene when the maneuver is being properly performed; 2. Introduction of the requirement to land at a particular spot, as opposed to the wide, flat open field initially used, and development of appropriate metrics of success; and 3. Inclusion of wind speed and wind direction settings (and random variability settings) to add a more realistic challenge in "hitting the spot."

Rogers, Steven P.↗

NASA NTRSDate not supplied

Airport Runway Configuration Management with Offline Model-free Reinforcement Learning

Runway configuration management (RCM) deals with the optimal selection of runways to operate on (for arrivals and departures) based on traffic, surface wind speed, wind direction and other environmental variables. RCM is one of the most challenging tasks in air traffic management, as it relies on operational and environmental variables (e.g., weather forecast) that are highly uncertain and complex to model. In this paper, an innovative and automated approach is deployed using offline model-free reinforcement learning to provide decision-support for RCM. The proposed technology processes historical data about variables of interest, decisions made regarding RCM, and their subsequent outcome, to identify a policy that would encourage good decisions and avoid the poor ones. The policy search is guided by an appropriately chosen weighted utility function (e.g., based on minimizing delays and go-arounds). Finally, the performance of the proposed tool is validated using Charlotte Douglas International Airport as the case study, which shows that the proposed method is superior to other conventional rule-based approaches.

Milad Memarzadeh↗

NASA NTRSDate not supplied

Airport Runway Configuration Management with Offline Model-free Reinforcement Learning

Runway configuration management (RCM) deals with the optimal selection of runways to operate on (for arrivals and departures) based on traffic, surface wind speed, wind direction and other environmental variables. RCM is one of the most challenging tasks in air traffic management, as it relies on operational and environmental variables (e.g., weather forecast) that are highly uncertain and complex to model. In this paper, an innovative and automated approach is deployed using offline model-free reinforcement learning to provide decision-support for RCM. The proposed technology processes historical data about variables of interest, decisions made regarding RCM, and their subsequent outcome, to identify a policy that would encourage good decisions and avoid the poor ones. The policy search is guided by an appropriately chosen weighted utility function (e.g., based on minimizing delays and go-arounds). Finally, the performance of the proposed tool is validated using Charlotte Douglas International Airport as the case study, which shows that the proposed method is superior to other conventional rule-based approaches.

Milad Memarzadeh↗