Prediction of weather impacted airport capacity using RUC-2 forecast

Abstract

Weather induced airborne delays have caused significant problems at the major commercial airports in the United States. Using Rapid Updated Cycle forecast data (RUC-2), this paper presents real case studies for assessing the impact of weather on airport capacities by Quadratic Response Surface (QRS) linear regression models and ensemble Bagging Decision Tree regression (BDT) models. Three highdemand major airports: Newark Liberty International Airport (EWR), Chicago O'Hare International Airport (ORD), and Atlanta International Airport (ATL) were selected for the analysis. Both QRS and BDT models were developed and evaluated using the Federal Aviation Administration (FAA) Aviation System Performance Metrics (ASPM) data and RUC -2 weather forecasts for these airports. The performances of the two weather impacted airport capacity models were compared and the errors between estimates and observations of airport arrival rates (AAR) were evaluated.The accuracy of AAR predictions in several hour look-ahead times using RUC-2 forecast weather was examined. The experimental results show that the accuracies of AAR estimations by nonlinear BDT regression models are much better than that from QRS multiple linear regression models. With airport AAR BDT models, the errors on weather corresponding AAR estimates using RUC weather data are smaller than that by METAR surface weather observations. Considering RUC includes much more weather information than METAR does, these findings strongly suggest that the adverse weather impact on airport capacity is complicated, relies on many predicting variables carried by RUC and it is inherently nonlinear. As a consequence, by using RUC weather forecasts, root mean squared error (RMSE) of the airport AAR BDT model predictions at one-hour look-ahead time are less than one and two arrival aircrafts per quarterly-hour for EWR and ORD/ATL airport, respectively. The ratio between AAR RMSE and average AAR for AAR predictions at one to eight hour look-ahead time is less than 10% for all three airports with nonlinear BDT models using RUC-2 weather forecasts. These results appear to be reasonably good.

Modelling the Impact of Adverse Weather on Airport Peak Service Rate with Machine Learning

“…Airport capacity modelling and prediction have been extensively addressed in the literature, employing diverse models and data sources. For instance, over a decade ago, the authors of [1,2] presented real case studies assessing the impact of weather on airport capacity using quadratic response surface linear regression and random forests. These models were trained using weather data from the Rapid Update Cycle (RUC) forecast.…”

Prediction of Hourly Airport Operational Throughput with a Multi-Branch Convolutional Neural Network

“…There have been extensive studies on modeling and predicting airport capacity by applying analytical methods [6][7][8][9] or artificial intelligence techniques [10][11][12][13][14]. The outcomes of those studies are expected to provide better inputs for air traffic flow management programs such as ground delay programs.…”

“…The comprehensive weather forecast product Rapid Cycle (RUC, now called Rapid Refresh or RAP) from the National Oceanic and Atmospheric Administration (NOAA) is another weather product used in aviation research. Although it was shown by Wang et al [11,12] that using surface weather elements from RUC to predict the airport acceptance rate (AAR) could achieve better accuracy than using METAR, RUC (now RAP) data have not been extensively used. Furthermore, RAP includes weather elements beyond the surface and covering different altitude ranges.…”