Impact of Weather Factors on Airport Arrival Rates: Application of Machine Learning in Air Transportation

Aug 4, 2026

Upload by Aurora

Weather is responsible for roughly 70% of air transportation delays in the National Airspace System, and convective-weather delays can impose millions of dollars in avoidable annual costs for airlines and passengers. The study investigates how environmental variables relate to airport efficiency estimates by mining archived weather records together with airport performance data from ten airports with distinct geography and climatology. Meaningful relationships were identified for six of the ten airports using multiple machine-learning methods, and the resulting models were validated with historical data.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Abstract

Introduction

Delay causes in the National Airspace System

Aircraft arrival rate as a key efficiency metric

Value of historical weather and performance databases

Prior research and modeling approaches

Motivation for predictive operational tools

Available online at [http://docs.lib.purdue.edu/jate](http://docs.lib.purdue.edu/jate) Journal of Aviation Technology and Engineering 12:2 (2023) 53–68 Impact of Weather Factors on Airport Arrival Rates: Application of Machine Learning in Air Transportation Robert W. Maxson 1, Dothang Truong2, and Woojin Choi2 1NOAA Aviation Weather Center 2Embry-Riddle Aeronautical University Abstract Weather is responsible for approximately 70% of air transportation delays in the National Airspace System, and delays resulting from convective weather alone cost airlines and passengers millions of dollars each year due to delays that could be avoided. This research sought to establish relationships between environmental variables and airport efficiency estimates by data mining archived weather and airport performance data at ten geographically and climatologically different airports. Several meaningful relationships were discovered from six out of ten airports using various machine learning methods within an overarching data mining protocol, and the developed models were tested using historical data. Keywords: data mining, airport arrival rate, flight delay, weather, machine learning I. Introduction The Federal Aviation Administration (FAA, 2015) outlines the major causes of delays in the National Airspace System (NAS) . These sources of delay (by the percentage of total delay) are attributed to weather (69%), traffic volume (19%), equipment failures (e.g., navigation, communications, surveillance equipment; 1%), runway unavailability (6%), and other miscellaneous causes (5%) . As documented by a review of NAS performance data collected over six years (from 2008 to 2013), adverse weather is the single largest cause of NAS delays, accounting for almost 70% of all delays (Sheth et al., 2015) . Delays generate enormous costs to both the flying public and airlines. In an FAA-sponsored National Center of Excellence for Aviation Operations Research (NEXTOR) report, Ball et al. (2010) estimated the total cost of flight delays in 2007 was $32 .9 billion. This estimate combined the direct costs borne by airlines and passengers as well as the more subtle indirect costs that ripple through the U.S. economy resulting from flight delays. In 2014, flight delay costs were estimated tobe $25 billion for U.S. air carriers by AviationFigure (2015) . As weather is responsible for the majority of flight delaysin the NAS (Sheth et al., 2015), a great deal of effort has been spent trying to predict and estimate the effects of weather on the NAS. 54 R. W. Maxson et al. / Journal of Aviation Technology and Engineering The key components necessary to enhance airspace efficiencies are accurate weather prediction and correctly converting these anticipated environmental conditions into expected impacts on scheduled traffic flows. A key metric in translating weather conditions and other impacts affecting air traffic flows at each major terminal is the aircraft arrival rate (AAR) . Per the FAA (2016), the AAR is‘‘a dynamic parameter specifying the number of arrival aircraft that an airport, in conjunction with terminal airspace, can accept under specific conditions throughout a consecutive sixty (60) minute period’’ (sec. 10-7-3) . FAA tactical operations managers along with terminal facility managers establish primary airport runway configurationsand associated AARs on at least a yearly basis for each facility or as required (e.g., as a result of airport construction or terminal airspace redesign) . The AAR establishes maximum airport capacity as a function of aircraft separation (miles-in-trail) on approach to the runway as determined by aircraft approach speeds. Based on a simple equation, average aircraft approach speeds (in knots) are divided by the desired miles-in-trail aircraft separation distance (with fractional remainders from this division conservatively rounded down to the nearest whole number) . It is fortunate that both the FAA

FAQs

Why does weather significantly affect airport operations and delays?

Weather accounts for about 69–70% of delay causes in the National Airspace System, making it the single largest contributor to delays. Convective weather in particular can lead to millions of dollars in annual avoidable costs.

What metric does the research use to connect weather with airport efficiency?

The study focuses on the Aircraft Arrival Rate (AAR), which represents the number of arrival aircraft an airport can accept under specific conditions over a consecutive 60-minute period.

How was machine learning applied and evaluated in the study?

The authors mined archived weather and airport performance data and tested multiple machine-learning methods within a data-mining protocol. Models were then evaluated using historical data, yielding meaningful relationships for six of ten airports.

Aug 4, 2026

Upload by Aurora