Nomenclature

APP

approach control

ATC

air traffic control

ATFM

air traffic flow management

ATCO

air traffic control officer

BWM

Best-Worst Method

CVM

Content Validity Method

CVR

Content Validity Ratio

DM

decision-maker

LTFM

Istanbul Airport

LTFJ

Istanbul Sabiha Gökçen Airport

LTAI

Antalya Airport

MTMA

military terminal areas

PMS

point merge system

RT

radio telephony

TMA

terminal control area

VL

validity levels

2.0 Literature review

Airspace complexity has long been a critical focus in air traffic management research. Laudeman et al. [Reference Laudeman, Shelden, Branstrom and Brasil3] introduced the concept of ‘dynamic density’, which combines aircraft count with the intricacies of traffic complexity. This pioneering concept shaped the development of ATC workload metrics and laid the groundwork for enhanced automation in air traffic management. Earlier, Pawlak et al. [Reference Pawlak, Brinton, Crouch and Lancaster4] emphasised the importance of evaluating both the cognitive and physical workloads of ATCOs, paving the way for a broader understanding of ATC operations.

Furthering this research, Histon et al. [Reference Histon, Hansman, Gottlieb, Kleinwaks, Yenson, Delahaye and Puechmorel5] investigated how airspace structure and procedural factors contribute to ATC cognitive complexity. They suggested that addressing these factors could help reduce overall complexity and ease the cognitive burden on ATCOs. Similarly, Riley et al. [Reference Riley, Chatterji, Johnson, Mogford, Kopardekar, Sieira, Landing and Lawton6] explored airspace complexity factors from the pilot’s perspective, focusing on variables such as the number of aircraft, climbing and descending trajectories, and speed variations. These findings underscore how such factors complicate conflict detection and resolution, significantly impacting operational decisions.

The study by Gianazza and Guittet [Reference Gianazza and Guittet7] marked a shift by prioritising specific complexity factors such as speed variations, aircraft count and potential conflicts, deeming these more critical than a broader set of factors. This focus on speed variations is particularly relevant in TMAs, where the continuous climb and descent of aircraft create additional challenges for ATCOs.

Kopardekar et al. [Reference Kopardekar, Bilimoria and Sridhar8] contributed substantially to this field with a real-time simulation study that quantified 52 different complexity factors, offering a scalable and nuanced metric for airspace complexity. While their work significantly advanced the understanding of these factors, achieving a universally accepted definition of airspace complexity – particularly from the perspective of ATCOs – remains an ongoing challenge. In another study, Kopardekar et al. [Reference Kopardekar, Prevot and Jastrzebski9] utilised factors such as the number of aircraft, the number of conflicts, aircraft density and vertical and horizontal convergence. They proposed conflict resolution using automation to reduce ATCOs’ efforts but left it to ATCOs to decide whether to implement it. These studies highlight the multifaceted nature of airspace complexity and the importance of ATCO judgement in managing complexity, which aligns with our focus on ATCO perceptions.

Brázdilová et al. [Reference Brázdilová, Cásek and Kubalčík10] approached complexity management by proposing changes in aircraft routes to distribute complexity more evenly across the airspace, aiming to create a more manageable environment. Prandini et al. [Reference Prandini, Putta and Hu11] focused on conflict and route optimisation in a four-dimensional context, highlighting the potential of strategic route adjustments in managing complex air traffic scenarios. Xiao et al. [Reference Xiao, Zhang, Cai and Cao12] emphasised the importance of identifying and analysing critical factors for complexity measurement, suggesting that not all factors contribute equally to complexity. By employing a genetic algorithm, they streamlined the complexity assessment process, focusing on key influencing factors.

Netjasov et al. [Reference Netjasov, Janić and Tošić13] introduced a comprehensive metric for terminal airspace complexity, incorporating static and dynamic factors for arrivals and departures. Their metric, validated using London Heathrow data, emphasised strategic planning rather than immediate workload management. More recently, Jurinić et al. [Reference Jurinić, Juričić, Antulov-Fantulin and Samardžić14] introduced new complexity indicators tailored to approach ATCO tasks, integrating operational challenges with airspace parameters to refine complexity management. An in-depth analysis of these studies reveals a wide array of factors contributing to airspace complexity. Table 1 summarises these factors alongside their references, providing a comprehensive overview of their role in previous research.

Table 1.

Complexity factors identified in the literature

Note: Table 1 summarises key complexity factors identified in previous studies, which are also considered in this research.

These factors are critical in understanding the challenges ATCOs face, particularly in TMAs like Istanbul, where traffic dynamics are highly complex. For instance, the ‘Number of Aircraft’ and ‘Possible Conflicts’ are particularly relevant in TMAs with heavy traffic, while ‘Performance Differences of Traffic’ and ‘Climbing/Descending Aircraft’ are significant in approach control due to varied aircraft types and flight phases.

While these studies provide a strong foundation for understanding airspace complexity, it is crucial to consider specific TMAs with unique operational challenges, such as Istanbul TMA. Istanbul TMA stands out due to its metroplex structure including multiple civil and military airports, diverse approach procedures and high traffic density, creating an exceptionally complex airspace environment. The need to manage varied procedures across multiple airports amplifies the workload and decision-making challenges faced by ATCOs. Understanding airspace complexity in such environments necessitates a deeper focus on ATCO perspectives, as their decisions are pivotal in ensuring safety and efficiency under these demanding conditions.

This study seeks to address this gap by focusing on the ATCO’s viewpoint, exploring their perceptions and the weighting they assign to various factors contributing to airspace complexity in a multi-airport-served TMA. This perspective aims to enrich the existing body of knowledge by integrating insights into the human factors that influence airspace management.

By synthesising the literature and incorporating ATCO perceptions, this research contributes to a more nuanced understanding of airspace complexity in approach control units. Through the application of a MCDM approach, this study not only identifies key complexity factors but also quantifies their relative importance as perceived by those directly managing the airspace.

3.0 Methodology

To comprehensively analyse the factors affecting airspace complexity within approach control units in the Istanbul TMA, this study adopts a mixed-method approach. By integrating qualitative and quantitative methodologies, we aim to capture a holistic understanding of complexity as perceived by experienced ATCOs. The methodology involves a case study of the Istanbul TMA, a critical and complex space managing both civil and military flights. The methodology is divided into two primary phases: the first phase employs the Content Validity Method (CVM) to identify relevant complexity factors, while the second phase uses the Best-Worst Method (BWM) to quantify the impact of these factors.

3.1 Case study: Istanbul TMA

The flight activity within Turkiye airspace from 2018 to 2023 exhibited a fluctuating trend. In 2018, the airspace saw 2.017.220 flights, with a marginal increase to 2.034.430 flights in 2019. The year 2020 witnessed a sharp decline to 1.055.168 flights due to the COVID-19 pandemic. A recovery was observed in 2021 with 1.466.860 flights. This upward trajectory continued in 2022, with the number of flights rising to 1.885.525. The year 2023 marked a significant rebound, reaching a total of 2.167.019 flights, indicating a return to pre-2020 levels and suggesting a resurgence in air travel within the region [21]. Providing ATC services with 75.245km of controlled flight routes and 982.286 square kilometers of controlled airspace, Turkiye has an important and large airspace where west-east and north-south air traffic flows between Europe, Asia, and the Middle East are carried out [22].

In terms of average daily movements, Istanbul Airport (LTFM) was the second busiest and Istanbul Sabiha Gökçen Airport (LTFJ) was the seventh busiest airport in Europe. In addition, these two airports are among the 30th busiest airports in European airspace with air traffic flow management (ATFM) delays for arrival traffic [21]. In 2022, LTFM and LTFJ had the highest amount of traffic in Turkiye [23].

LTFM and LTFJ, which are among the busiest airports in Europe, and Turkiye, and the approach control service for these airports is provided by Istanbul TMA. Istanbul TMA provides ATC services to seven civil airports (LTFM, LTFJ, LTBA, LTBW, LTBU, LTBR and LTBQ) and three military airports (LTBX, LTBG, and LTBP) as seen in Fig. 1. In addition, there are 24 training areas, 2 military terminal areas (MTMAs), and many restricted/prohibited airspaces. Although the point merge system (PMS) is implemented at both LTFM and LTFJ, other airports within service areas of the same approach unit utilise traditional procedures, highlighting the coexistence of different operational methods within a single airspace.

Figure 1.

Map of Istanbul TMA including multiple civil and military airports [24].

(Note: Figure 1 depicts the geographical layout of Istanbul TMA, highlighting the proximity of multiple airports and special-use airspaces.)

Figure 2 illustrates the yearly air traffic trends for airports in the Istanbul TMA from 2018 to 2023. A significant event occurred in 2018 with the opening of LTFM. This resulted in a sharp decline in traffic at Atatürk Airport (LTBA) as operations gradually shifted to the new airport. This trend is visible as traffic increased at LTFM while simultaneously decreasing at LTBA.

Figure 2.

Total aircraft movements in Istanbul TMA from 2018 to 2023 (take-off and land).

Between 2019 and 2020, a notable drop in traffic can be attributed to the global COVID-19 pandemic, which severely impacted air travel worldwide. Post-pandemic recovery is evident in subsequent years, with LTFM and LTFJ emerging as the two busiest airports in the region. As of 2023, these two airports handle the majority of the air traffic within the Istanbul TMA. These key transitions and trends highlight the dynamic nature of air traffic management in one of the world’s most strategically located aviation hubs.

Approximately 45% of the traffic taking off from or landing in Turkish airspace operates from one of the airports within the Istanbul TMA. Around 94% of traffic landing or departing from airports in the Istanbul TMA did so at one of Istanbul’s airports. Given that nearly 42% of landings and departures in Turkiye’s airspace are to Istanbul city, the Istanbul TMA is the busiest TMA in the country [23] (Fig. 3).

Figure 3.

Total aircraft movements in Turkiye and Istanbul TMA from 2018 to 2023 (take-off and land).

3.2 Data analysis

This study adopts a mixed-method approach to comprehensively analyse factors affecting airspace complexity, particularly in a TMA within approach control units. After a thorough literature review, a pool of complexity metrics was generated, from which 50 metrics were selected for initial assessment by combining and eliminating similar or overlapping factors. Then the methodology is structured into two primary phases: first, identifying the factors influencing TMA complexity through a qualitative approach, and second, quantifying the impact of these factors using quantitative methods.

3.2.1 Qualitative analysis using CVM

The CVM process begins with creating an initial draft of an assessment tool and relies on a group of experts to evaluate and rate its features based on how well they represent and are relevant to the content area [Reference Lynn25]. A total of 40 ATCOs, all with at least three years of experience in managing approach control operations, were selected as the first working group to provide insights on various factors contributing to airspace complexity. The CVM was employed to ensure the identified factors were relevant and significant from the ATCOs’ perspective. The ATCOs were presented with 50 complexity factors derived from the literature and asked to classify each as “essential,” “useful but not essential” or “not necessary.”

The content validity ratio (CVR) for each factor was calculated using Lawshe’s formula:

\begin{align}CVR = \frac{{{N_E} - \left( {N/2} \right)}}{{N/2}}\end{align}

Where:

• • ${N_E}$ is the number of ATCOs rating the factor as essential.

• • $N$ is the total number of ATCOs.

These CVR values were then compared against Lawshe’s minimum values given in Table 2 to determine the essentiality of each factor [Reference Lawshe26].

Table 2.

Minimum values of CVR

According to Lawshe’s table, for 40 participants, the minimum CVR value required for a factor to be considered essential is 0.29. Factors with CVR values above this threshold were included in the subsequent analysis. For example, a complexity factor that was deemed essential by only 25 out of 40 ATCOs resulted in a CVR of 0.25. This value indicates that the factor fails to meet Lawshe’s minimum threshold of 0.29 for essentiality; therefore, these values have not been included in our assessment.

3.2.2 Quantitative analysis using Best-Worst Method (BWM)

When dealing with multiple factors, it’s necessary to use MCDM methods to compare these factors and determine their relative importance. In this study, the BWM developed by Rezaei [Reference Rezaei27] is preferred for its efficiency in using fewer pairwise comparisons than other MCDM methods and for providing more accurate data on the structure of the pairwise comparison system [Reference Salimi and Rezaei28].

The second phase of the study involves analysing the factors identified in the first phase with the BWM to determine the extent of their impact on airspace complexity. Given the number of factors to be evaluated, the BWM was structured in two stages. Due to the complexity of comparing more than nine criteria, it is recommended to divide the criteria into main and sub-groupings [Reference Rezaei27]. The main and sub-grouping of the criteria were carried out with ATCOs, who are specialists in the field of air traffic control.

In this context, the BWM questionnaire was administered to 20 ATCOs with a minimum of ten years of experience working at İstanbul Atatürk Airport approach control unit (which serves all of Istanbul TMA). The extensive presence of all sub-criteria in Istanbul airspace allowed for precise evaluation. Participants with pairwise comparison validity levels (VL) above the acceptable threshold were asked to review their comparisons, ensuring that all pairwise comparison validity levels were acceptable. The average validity level of each pairwise comparison for each criterion is shown in Table 3.

Table 3.

CVR average validity levels for criteria in best-worst method analysis

Description: Table 3 presents the average validity levels for each criterion as assessed through the BWM. Validity levels reflect the consistency of pairwise comparisons, where a lower validity level indicates higher consistency.

Following the BWM, an arithmetic mean approach was used to evaluate the values received from each ATCO’s viewpoint, determining the weights of the criteria and sub-criteria.

The process of comparing criteria with BWM involves several steps:

1. 1. Identification of Criteria: Determine the criteria to be used in BWM.

(1) \begin{align} Criteria:\left\{ {{C_1},{C_2}, \ldots, {C_n}} \right\} \end{align}

Here, ${C_1}$ represents the first criterion, ${C_2}$ the second, and ${C_n}$ the $n$ -th criterion.

1. 2. Identifying the Best and Worst Criteria: The decision-maker (DM) selects the best and the worst criteria among the set.

2. 3. Comparison of the Best Criterion Against All Others: The DM compares the best criterion with all others based on an importance scale of 1-9, resulting in vector ${A_B}$ :

(2) \begin{align}{A_B} = \left\{ {{a_{B1}},{a_{B2}}, \ldots, {a_{Bn}}} \right\}\end{align}

Here, ${a_{Bj}}$ indicates the importance of the best criterion relative to the $j$ -th criterion.

1. 4. Comparison of All Criteria Against the Worst Criterion: The DM compares all criteria against the worst criterion based on the importance scale, resulting in vector ${A_W}$ :

(3) \begin{align}{A_W} = \left\{ {{a_{1W}},{a_{2W}}, \ldots, {a_{nW}}} \right\}\end{align}

1. 5. Calculation of Criterion Weights: The optimal weight for each criterion is calculated. The ideal situation is where $\frac{{{w_B}}}{{{w_j}}} = {a_{Bj}}$ and $\frac{{{w_j}}}{{{w_W}}} = {a_{jW}}$ for each pair. The problem is formulated as a linear programming problem to minimise the maximum deviation among these ratios:

(4) \begin{align}\min {\xi _L}\end{align}

Subject to:

(5) \begin{align}\left| {{w_{B}} - {a_{Bj}}{w_j}} \right| \le {\xi _L},\quad \forall j\end{align}

(6) \begin{align}\left| {{w_j} - {a_{jW}}{w_W}} \right| \le {\xi _L},{\rm{\;\;\;\;}}\forall j\end{align}

(7) \begin{align}\mathop \sum \nolimits_{j = 1}^n {w_j} = 1\end{align}

(8) \begin{align}{w_j} \geqslant 0,\quad \forall j\end{align}

The optimal weights $\left( {w_1^{\rm{*}},w_2^{\rm{*}}, \ldots, w_n^{\rm{*}}} \right)$ and $\xi _L^{\rm{*}}$ are obtained after solving this problem, with $\xi _L^{\rm{*}}$ indicating the consistency of the comparison method. A lower value of $\xi _L^{\rm{*}}$ signifies better consistency and more reliable comparisons.

Integrating the results of CVM and BWM provides a comprehensive understanding of complexity factors in TMAs. The qualitative insights from CVM assist in selecting factors for BWM analysis, ensuring that only relevant complexity factors are considered based on expert judgement. Additionally, BWM’s quantitative outcomes assign weights to these factors, reflecting their significance as perceived by experienced ATCOs. This dual approach not only enhances the rigor of the analysis but also incorporates both qualitative insights and quantitative data related to airspace complexity, allowing for a more comprehensive evaluation. Furthermore, the gap between theoretical frameworks and practical applications is bridged by combining qualitative insights with quantitative data, facilitating better decision-making in air traffic management.

5.0 Discussion and conclusion

This study has demonstrated a comprehensive approach to analysing airspace complexity in approach control units, particularly in metroplex airspaces, using the Istanbul TMA as a case study. By employing the CVM and BWM methods, the research identified and quantified the critical factors influencing airspace complexity as perceived by ATCOs. The findings emphasise the pivotal role of ‘Conflicts,’ ‘Airspace’ and ‘Traffic Density,’ highlighting the importance of conflict resolution, efficient airspace design and traffic management in ATC operations. Moreover, detailed sub-criteria weights reveal operational challenges such as crossing conflicts, procedural constraints and workload distribution.

The practical applications of these findings within the ATM system are significant. Airspace planners can use the criteria and sub-criteria weights to prioritise improvements, focusing resources on the most impactful factors. For example, decision-support tools can be developed to assist ATCOs in managing complexity more effectively. Weighted factors can also guide the redesign of air routes to minimise crossing conflicts or the revision of procedures to increase operational flexibility.

This research underscores the need to contextualise airspace complexity within specific operational environments, such as Istanbul TMA. The correlation values presented offer valuable insights into how different factors interact, aiding decision-makers in identifying key areas for intervention. However, it is essential to recognise that these interactions are not universally fixed but vary across traffic characteristics, organisational practices and airspace configurations. Future research should explore these variations to develop more adaptable and widely applicable complexity indicators.

In the context of capacity improvement and workload reduction in Istanbul TMA, the study highlights the need to focus on conflict management, procedural streamlining and airspace optimisation. By addressing these key areas, ATCOs can enhance efficiency, potentially increasing airspace capacity while reducing workload. For instance, redesigning air routes to alleviate crossing conflicts or adapting procedures for greater flexibility can significantly improve operations. This study provides airspace planners, policymakers and ATCOs with actionable insights to better manage the complexities of modern air traffic systems.

6.0 Limitations and future work

This study primarily draws upon data from the Istanbul Atatürk Airport approach control unit, reflecting specific operational and environmental conditions. These include the presence of multiple civil and military airports, the use of advanced systems like PMS and high traffic volumes. While these factors enrich the study as a detailed case, they also limit the generalisability of the findings to other airports with differing traffic patterns, airspace structures or operational constraints. Airports in diverse regions with varying sizes, traffic compositions or procedural standards may exhibit distinct complexities and ATCO experiences. Therefore, caution is necessary when applying these findings to other airspace environments.

Airspace complexity is inherently dynamic, shaped by evolving traffic patterns, technological advancements, and changing air traffic management practices. This study captures ATCO perceptions at a specific point in time; however, these perceptions may shift with the introduction of new technologies, changes in operational practices, or as controllers gain additional experience. Consequently, the findings may not fully reflect the long-term evolution of airspace complexity.

Certain factors, such as military traffic, environmental protection (noise abatement) and runway configurations, were excluded from the final analysis. While relevant, these factors are either managed through separate procedures or are encapsulated within broader criteria like ‘Airspace’ and ‘Procedures.’ For instance, noise abatement is considered a sub-component of procedural constraints, while runway configurations are factored into flow organisation.

Future research can expand upon these findings by exploring broader geographical regions, incorporating objective complexity measures, and examining the temporal dynamics of complexity. Advanced technologies, such as AI-based conflict detection and resolution tools, dynamic sectorisation models, and decision-support systems, could further enhance air traffic management. For example, real-time monitoring tools for workload distribution and adaptive coordination mechanisms for multi airport TMAs could improve operational efficiency and safety.

In conclusion, this study lays a foundation for understanding airspace complexity and offers actionable insights for future air traffic management strategies. By addressing key complexity factors, such as conflict management, traffic density and procedural constraints, future initiatives can create safer, more efficient and adaptable airspace environments.