Priorities, interdependencies, and pathways for net-zero energy building adoption in the UAE: a Bayesian network-informed mixed-methods study

1 Introduction

The building sector is a major source of global energy demand and greenhouse-gas emissions, making its decarbonization central to climate-mitigation efforts (Ohene et al., 2022; Þórólfsdóttir et al., 2023). Net-Zero Energy Buildings (NZEBs) address this challenge by first reducing operational energy demand and then balancing the remaining demand with on-site or nearby renewable-energy generation (Ahmed et al., 2021; Sartori et al., 2012). Their value extends beyond energy performance to energy security, climate resilience, and sustainable urban development. Demonstration projects further show that integrated design, renewable generation, and performance monitoring can move buildings beyond net-zero toward positive-energy operation (Barrutieta et al., 2023).

Technical feasibility alone does not determine widespread NZEB adoption. Regulation, financing, professional capability, stakeholder coordination, market readiness, and user behavior shape how high-performance designs are approved, delivered, commissioned, and operated (Ohene et al., 2022; Mohammed et al., 2023; Mavrigiannaki et al., 2021; Godin et al., 2021). NZEB adoption is therefore a socio-technical implementation endeavor rather than a purely engineering one. This distinction is particularly important in hot-arid settings, where substantial cooling requirements, rapid urban development, renewable-energy integration, and long-term operational performance must be addressed together.

The United Arab Emirates (UAE) provides a strategically important and forward-looking setting for examining how national sustainability ambition can be translated into wider NZEB implementation. The UAE Energy Strategy 2050, the Net Zero by 2050 Strategic Initiative, Masdar City, and Dubai’s Clean Energy Strategy 2050 demonstrate sustained national commitment. Abu Dhabi’s Estidama Pearl Rating System has required minimum sustainability performance for new buildings since 2010 (Awadh, 2018). Dubai Electricity and Water Authority (DEWA) introduced Shams Dubai in 2015, enabling building owners to generate electricity through on-site photovoltaic systems and export surplus electricity to the grid for bill credit (Dubai Electricity and Water Authority, 2024). The Emirates Green Building Council subsequently established the Net Zero Centre of Excellence to support policy development, technical readiness, and industry engagement (Emirates Green Building Council, 2024; Asfour and Mansour, 2026). Taken together, these national and emirate-level initiatives demonstrate an established enabling architecture spanning national strategy, green-building performance, distributed solar participation, and industry capability development (Awadh, 2018; Asfour and Mansour, 2026).

These initiatives, together with the country’s exceptional solar potential and continuing urban development, create a strong foundation for climate-responsive NZEB implementation (Pandita et al., 2024). The analytical opportunity is to examine how these established strengths can be connected and scaled across project delivery under the UAE’s distinctive climatic and market conditions. The hot-arid environment makes advanced passive design, high-efficiency cooling, and solar integration particularly valuable; retrofit and system-integration requirements create space for targeted innovation; and financing, technical capability, and public engagement can reinforce performance across the building lifecycle (Þórólfsdóttir et al., 2023; Ahmed et al., 2021; Fayyad and John, 2017). Building-integrated photovoltaics (BIPV), energy storage, artificial-intelligence-driven energy management, and smart-grid integration further expand the country’s implementation options.

Existing NZEB research is extensive in energy efficiency, life-cycle assessment, building simulation, and renewable-energy integration, while stakeholder engagement, neighborhood-scale systems, and climate-responsive adoption in arid and tropical regions remain less developed (Ohene et al., 2022; Þórólfsdóttir et al., 2023). Recent reviews have strengthened the evidence on passive-active design, flexible energy systems, and implementation conditions (Brown et al., 2024; Lu et al., 2024). Nevertheless, three streams remain insufficiently connected: technical design evidence, stakeholder priority evidence, and system-interdependence evidence. Adoption studies commonly rank individual implementation factors independently, but provide less insight into whether highly rated factors also occupy influential positions within the wider implementation system. This matters because a priority list identifies what stakeholders regard as important, whereas an interdependence model can reveal how those factors are positioned within a connected adoption structure.

The research gap lies in the literature’s limited integration of two decision perspectives: which implementation challenges and enablers stakeholders prioritize when each factor is assessed independently, and which factors occupy influential positions within a modeled dependency network. The scientific novelty lies in comparing those two orderings within one mixed-methods design and using their convergence or divergence to derive a coordinated intervention portfolio rather than another undifferentiated factor list. The UAE offers a strong application context for this comparison because its national strategies, green-building requirements, and technical-readiness initiatives already provide a substantive platform for analysis and implementation (Emirates Green Building Council, 2024; Asfour and Mansour, 2026; Fayyad and John, 2017). Accordingly, the study examines how an active sustainability ecosystem can be further coordinated to support wider NZEB delivery.

Against this background, the study aims to explain the priority and interdependence structures shaping NZEB adoption in the UAE and to translate them into actionable pathways. Three objectives operationalize this aim: (1) to identify and independently prioritize context-specific implementation challenges and enablers; (2) to model their conditional interdependencies and compare modeled network propagation with standalone priority; and (3) to derive a coordinated portfolio of policy and practice responses. These objectives are expressed through three research questions: RQ1. Which implementation challenges and enablers do UAE stakeholders identify and prioritize? RQ2. How do those independent priorities compare with the factors that occupy influential positions in a Bayesian Belief Network (BBN) of conditional interdependencies? RQ3. Which coordinated policy and practice responses follow from the combined evidence?

The study uses a sequential exploratory mixed-methods design to answer these questions. Expert interviews identify and contextualize the factor set, the Likelihood-Impact Relative Importance Index (LI-RII) ranks each factor independently on a normalized scale, and the BBN represents modeled conditional dependencies and produces Network Propagation Impact (NPI) scores. The design supports three contributions. Empirically, it provides context-specific evidence on NZEB adoption in a hot-arid and rapidly developing setting. Methodologically, it separates normalized standalone priority from model-derived network propagation, compares rank positions rather than incompatible raw scores, and quantifies rank correspondence through Spearman and Kendall statistics. Practically, it translates the combined evidence into a coordinated implementation portfolio spanning regulation, finance, technical integration, workforce capability, and public engagement.

The remainder of the paper is organized as follows. Section 2 develops the conceptual basis and establishes the research gap. Section 3 explains the sequential mixed-methods design and the separate roles of LI-RII and the BBN. Section 4 presents the qualitative findings, LI-RII results, BBN results, and the structured comparison of both rankings. Section 5 interprets the areas of agreement and divergence and develops the study’s theoretical and practical implications. Section 6 concludes by closing the research questions, acknowledging the study’s limitations, and outlining directions for future research.

2 Literature review

The NZEB literature has evolved from defining energy balance and technical design requirements to examining how high-performance buildings can be financed, delivered, operated, and scaled. This broader implementation perspective is essential because net-zero performance depends not only on reducing demand and supplying renewable energy, but also on the institutional, market, professional, and behavioral conditions that allow those measures to function throughout the building lifecycle (Ohene et al., 2022; Þórólfsdóttir et al., 2023; Ahmed et al., 2021; Sartori et al., 2012).

This review is therefore organized around the logic of adoption rather than as a catalogue of technologies. It first clarifies NZEB definitions and standards, then examines design and performance strategies, implementation conditions at component, project, and system levels, transferable international experience, and emerging smart technologies. Together, these five themes establish what NZEB delivery requires, which conditions shape wider adoption, and why a UAE-specific analysis must connect technical evidence with stakeholder priorities and modeled interdependencies.

2.1 Definitions and standards

A consistent definition of NZEBs is necessary for regulatory alignment, performance benchmarking, design decisions, and cross-project comparison. Sartori et al. (2012) distinguish NZEBs from broader green-building concepts by requiring a measurable balance between energy consumed and renewable energy produced over a defined period, commonly 1 year. Green buildings may address a wider set of environmental objectives, whereas NZEBs specifically prioritize deep operational-demand reduction and renewable-energy compensation of the remaining demand.

Related concepts represent different levels of building energy performance (Þórólfsdóttir et al., 2023; Sartori et al., 2012; Wu and Skye, 2021). Nearly zero-energy buildings achieve very low demand but may continue to rely partly on external energy supply. NZEBs balance annual consumption with renewable generation, while net-plus energy buildings generate a net surplus. Recent European evidence also shows movement from nearly zero-energy requirements toward zero-emission building standards (Maduta et al., 2025). The coexistence of these concepts reinforces the need for definitions that are measurable, transparent, and adaptable to different climatic and regulatory contexts.

Sartori et al. (2012) identify six elements for defining and evaluating net-zero performance: weighted energy demand, weighted energy supply, reference-building comparison, the net-zero balance line, energy-efficiency requirements, and renewable-energy supply requirements. These elements clarify the accounting boundary, benchmark, and relationship between demand reduction and renewable supply. They provide the technical foundation for comparing projects without assuming that one design solution or national standard is universally applicable.

Climate-responsive NZEB design is especially important in the UAE. The country’s hot-arid climate, substantial cooling requirements, and exceptional solar resource shape annual energy profiles and create strong opportunities for integrated demand reduction and renewable generation. NZEB benchmarks for the UAE should therefore combine rigorous demand reduction with climate-responsive envelopes, efficient cooling, and realistic renewable-energy accounting. Locally relevant definitions support consistent application of net-zero principles under the UAE’s operating conditions.

2.2 Design and performance strategies

NZEB performance requires integrated passive and active design rather than isolated technology selection. Wu and Skye (2021) identify airtight and well-insulated envelopes, appropriate orientation, efficient heating, ventilation, and air-conditioning (HVAC) systems and lighting, on-site renewable generation, storage and demand response, and continuous monitoring and commissioning as central elements. Recent reviews reinforce this systems logic through coordinated passive-active design, flexible energy-system planning under uncertainty, and climate-specific implementation (Mistarihi et al., 2025; Maghami et al., 2025; Lu et al., 2024). Evidence from a high-performance office building also demonstrates the importance of daylighting integration and performance monitoring (Guglielmetti et al., 2011). For the UAE, passive cooling, solar optimization, efficient cooling systems, and intelligent energy management must therefore be designed as one operating system.

The UAE Net Zero by 2050 Strategic Initiative and related programs provide a strong foundation for scalable NZEB practice (Emirates Green Building Council, 2024; Asfour and Mansour, 2026; Fayyad and John, 2017). Table 1 distinguishes the elements that determine the NZEB energy balance from complementary sustainability criteria that contribute to high-performance buildings in the UAE.

Table 1. Qualitative assessment criteria for NZEBs.

Table 1 clarifies the conceptual boundary of the study. Energy efficiency, smart energy management, and renewable-energy integration directly determine the net-zero balance. Water conservation, indoor environmental quality, sustainable materials, and site considerations do not independently establish NZEB status, but they affect environmental quality, user acceptance, durability, and lifecycle performance. These dimensions therefore served as sensitizing categories for the interview guide and technical discussion; they were not treated as the final survey factor set.

2.3 Adoption conditions and implementation considerations

The implementation literature can be organized at three analytical levels. Table 2 examines building components and technical systems, Table 3 addresses materials and supporting infrastructure, and Table 4 summarizes cross-cutting financial, regulatory, coordination, and behavioral conditions. This layered structure is important because component-level considerations can reflect project delivery or wider institutional arrangements rather than the technology alone. For the UAE, this international evidence provides a comparative background for examining how existing policy, technical, market, and infrastructure strengths can support wider NZEB delivery.

Table 2. Challenges in implementing NZEB components.

Table 3. Challenges in sustainable materials and site selection.

Table 4. Systemic and cross-cutting barriers to NZEB adoption.

At the component level, the UAE’s advanced construction sector, exceptional solar potential, and hot-arid climate create a compelling context for high-performance envelopes, efficient HVAC systems, smart controls, and renewable-energy technologies (Mohammed et al., 2023; Brown et al., 2024; Mistarihi et al., 2025; Wu and Skye, 2021). The country’s established technical and institutional capacity provides a solid basis for managing initial investment and sustaining specialist design, installation, commissioning, and maintenance capability. Smart-energy deployment is further supported by interoperability, robust data governance, and user-centered controls (Mohammed et al., 2023; Mavrigiannaki et al., 2021). Recent systematic and UAE-focused evidence reinforces the value of coordinating cost, policy alignment, knowledge, workforce capacity, and technical integration as mutually supporting implementation dimensions (Brown et al., 2024; Mistarihi et al., 2025).

Table 2 identifies the main implementation conditions associated with advanced NZEB components. Upfront investment, technical complexity, commissioning, and maintenance affect high-performance envelopes, HVAC, smart systems, and renewable technologies. The UAE’s hot-arid operating environment and strong solar resource further strengthen the case for climate-responsive design, robust equipment selection, and specialized operation. These conditions create opportunities for continued professional development, cost-effective innovation, and systems that integrate smoothly with the country’s rapidly evolving building infrastructure.

Material supply and supporting infrastructure extend the implementation system beyond individual building technologies. As shown in Table 3, stronger local supply chains for recycled or low-carbon materials and expanded recycling capacity can support cost efficiency and circular construction. Access to low-carbon transport and supporting electrical infrastructure does not determine the net-zero energy balance directly, but it influences project feasibility and the wider sustainability value of NZEB development. These factors are therefore implementation conditions, not alternative definitions of NZEB performance.

Table 3 broadens the analysis from building components to project and urban context. Material availability, construction-waste practices, transport integration, walkability, and charging infrastructure can shape delivery decisions and the wider environmental contribution of high-performance projects. Their analytical role is to explain the conditions surrounding NZEB implementation while preserving energy balance as the defining performance criterion.

Beyond technology and infrastructure, wider adoption is supported by institutional and behavioral conditions. Table 4 summarizes complementary areas through which the UAE’s existing sustainability momentum can continue to expand, including public and industry awareness, accessible financing, ongoing regulatory development, and coordination among technical, planning, and policy actors. The UAE has already established ambitious national targets and green-building requirements. The central opportunity is to extend alignment across emirates, project types, and stages of the building lifecycle so that leading practice can diffuse more widely. Accordingly, these factors are interpreted as opportunities to amplify and connect mechanisms already established through the UAE’s sustainability strategies and green-building programs (Awadh, 2018; Dubai Electricity and Water Authority, 2024; Emirates Green Building Council, 2024; Asfour and Mansour, 2026; Fayyad and John, 2017).

The factors in Table 4 show how financial, institutional, and behavioral support complements technical readiness. Broader awareness can strengthen demand and informed operation; financing can improve project feasibility; regulatory clarity can support efficient decision-making; and interdisciplinary coordination can connect policy intent with design and delivery. These mechanisms build directly on the UAE’s existing progress and provide a practical basis for scaling NZEB principles across an already active sustainability ecosystem.

Taken together, Tables 2–4 show that NZEB adoption is a multilevel implementation system spanning technology, projects, institutions, markets, and users. They supplied the initial domains used to structure the expert interviews, while the final factor set was refined through UAE stakeholder evidence before inclusion in the survey. The central analytical question is which conditions receive the greatest standalone priority and which occupy the widest positions within a connected adoption system.

2.4 Globally transferable insights

International evidence provides adaptable mechanisms that can complement the UAE’s established sustainability framework. The European Union (EU) Horizon 2020 ZERO-PLUS project showed that residential communities can approach net-zero performance through integrated urban planning, shared renewable-energy systems, and coordinated energy management across multiple buildings (Mavrigiannaki et al., 2021). This experience demonstrates the value of moving beyond individual projects toward district- and community-scale energy coordination.

Evidence from Egypt, Canada, and Sub-Saharan Africa likewise shows that NZEB implementation depends on governance, financing, technical capability, and stakeholder participation (Mohammed et al., 2023; Godin et al., 2021; Omar et al., 2022). Public awareness and financial support can strengthen demand and offset the initial costs of efficient technologies and renewable-energy systems. Across these contexts, the common lesson is that technical solutions scale when institutional, financial, and professional conditions develop alongside them.

Regulatory pathways also vary with policy priorities, market capacity, and climate. European evidence documents a transition from nearly zero-energy requirements toward zero-emission standards while retaining national variation in definitions and thresholds (Maduta et al., 2025). Canada combines public investment with phased regulatory enforcement (City of Toronto, Environment and Climate Division, 2023). Jordan demonstrates the relevance of passive cooling and photovoltaic integration in a hot-arid setting (Ali et al., 2020). These cases support adaptation rather than direct transfer of international models.

The international literature therefore contributes two insights. First, successful adoption usually combines regulatory direction, financing, technical capability, and stakeholder participation. Second, the relative importance and sequencing of these mechanisms remain context dependent. In this study, global evidence defines candidate enablers and adaptation principles, while UAE stakeholders determine local relevance and priority.

2.5 Emerging smart NZEB technologies

Smart NZEB technologies extend the implementation agenda from efficient design to adaptive operation. Digital controls, post-occupancy feedback, and connected energy systems can improve performance, while their value is reinforced by interoperability, commissioning, skilled operation, and context-sensitive system design (Mistarihi et al., 2025; Maghami et al., 2025; Lu et al., 2024; Karlessi et al., 2017; Song et al., 2014). Occupancy-responsive HVAC and lighting, smart-grid connectivity, peak-load shifting, real-time monitoring, and artificial-intelligence-driven optimization therefore increase both technical opportunity and the importance of integration and user acceptance.

For the UAE, smart operation is particularly relevant because solar availability is high and cooling requirements vary with season, occupancy, and building use. Adaptive controls and connected energy systems can reduce peak loads, support demand-side management, and improve resilience and user experience (Karlessi et al., 2017; Song et al., 2014). Performance-gap research shows that outcomes depend on lifecycle information, control strategies, management practices, and occupant interaction as well as design intent (Zheng et al., 2024; Mahdavi and Berger, 2025). The decision-design-feedback logic proposed by Song et al. (2014) is therefore valuable because it connects design choices with operational evidence and continual adjustment. These characteristics position the UAE well for advanced operational strategies that combine renewable generation, intelligent controls, and demand-side flexibility.

Across the literature, five areas for continued acceleration emerge for the UAE: continued refinement of clear and measurable NZEB performance rules; deeper coordination among technical, planning, and policy actors; further alignment of financing and incentives; continued workforce and public-capacity development; and climate-responsive adaptation of international practice. These areas build directly on the country’s existing policy and technical foundations and are consistent with recent systematic and UAE-focused evidence (Brown et al., 2024; Mistarihi et al., 2025). They are generally presented as parallel recommendations, with limited evidence on whether their importance changes when interdependencies are modeled.

The review identifies three connected gaps: the separation of technical-performance and adoption research, the limited treatment of dependencies in factor-prioritization studies, and the need for context-sensitive evidence from hot-arid, rapidly developing settings. Together, these gaps motivate an analysis that links stakeholder judgments with a connected view of NZEB implementation.

The literature review leads directly to the empirical design. Definitions and design studies establish what NZEB performance requires; implementation studies identify the technical, economic, regulatory, organizational, and behavioral conditions that shape delivery; and international cases provide candidate intervention mechanisms. The expert interviews test and refine these domains for the UAE. The survey and LI-RII answer the standalone-priority component of RQ1, the BBN and NPI analysis address modeled interdependencies under RQ2, and the integrated comparison provides the evidence base for the coordinated pathways developed under RQ3.

3 Materials and methods

This study adopted a sequential exploratory mixed-methods design because the research problem required both context-specific factor development and a structured comparison of standalone priority with modeled interdependence. The design contained two explicit integration points. First, broad technical, economic, regulatory, and behavioral domains identified in the literature were refined through expert interviews into the final set of challenges and enablers. Second, LI-RII and BBN outputs derived from the same survey evidence were compared during interpretation. This sequence aligned the methods with the research questions: factor identification and independent prioritization for RQ1, comparison with modeled network propagation for RQ2, and synthesis into coordinated policy and practice pathways for RQ3.

Ethics and consent. The Institutional Review Board (IRB) of the American University of Sharjah (AUS) granted Exempt approval under 45 CFR 46.104(d)(2) for Protocol 25–009, “A strategic framework of Net-Zero Energy Buildings in the UAE” (IRB review date: November 24, 2024; effective date: November 25, 2024). The approval correspondence confirms that AUS-approved informed-consent language accompanied the approved protocol. The study was conducted within that approved consent framework, and participant identities are not reported in the article.

• Qualitative phase. Five face-to-face semi-structured interviews, each lasting approximately 30 to 45 min, were conducted with mid-to-senior professionals recruited through purposive and snowball sampling. Participants represented utilities, consultancy, engineering, and academia and had 19 to 31 years of professional experience. The interview guide covered renewable-energy integration, regulation, finance, technological options, retrofitting, and public acceptance. One researcher conducted the thematic analysis using the following sequence: verbatim transcription, repeated transcript familiarization, initial coding, grouping of related codes into candidate themes, review of candidate themes against the transcripts, and final theme definition. Rigor was supported through iterative transcript checking and explicit traceability from themes to the final factor set. The adequacy of five interviews was assessed through information power, reflecting the narrow contextual aim, highly specific expert sample, established literature base, and in-depth professional dialogue (Guest et al., 2006; Braun and Clarke, 2021; Palinkas et al., 2015; Malterud et al., 2016). The interviews therefore provided focused contextual evidence for factor refinement.
• Quantitative phase. The literature-interview synthesis was translated into a structured survey. Before full administration, the instrument was pilot-tested with three individuals. The final survey comprised 53 items: five professional-profile items and 48 factor ratings, with separate likelihood and impact ratings for 15 challenges and nine enablers. Both dimensions were measured on five-point scales. Instrument content was established through direct traceability from the literature domains and interview themes to every retained factor, followed by separate likelihood and impact wording for each factor. This traceability-based procedure supports contextual content coverage and aligns the instrument with its purpose as a context-specific prioritization tool. Thirty usable responses were analyzed. The BBN was estimated as two separate exploratory networks, comprising 15 challenge nodes and nine enabler nodes. Accordingly, the quantitative phase is interpreted as exploratory prioritization, conditional-dependency structure, and comparative decision support.

Step 1: LI-RII ranking. The LI-RII expressed the standalone priority of each challenge and enabler from paired likelihood and impact ratings. The approach follows risk-matrix logic used to combine the two decision dimensions in construction risk analysis (Qazi and Dikmen, 2021) and addresses the prioritization component of RQ1 before modeled interdependencies are considered.

For each factor, the LI-RII was calculated as follows:

LI − RII j = 1 25 N ∑ i = 1 N ( L ij × I ij ) (1)

For respondent i and factor j, the likelihood rating Lᵢj and impact rating Iᵢj were multiplied to obtain the influence score Sᵢj = Lᵢj × Iᵢj. Each rating ranged from 1 to 5, so the product ranged from 1 to 25. In Equation 1, N denotes the number of respondents, and division by 25 N normalizes the mean influence score. The attainable LI-RII range is therefore 0.04 to 1.00. Multiplication gives greatest priority to factors judged both likely and consequential, consistent with the study’s risk-matrix-based screening purpose. This use of the ratings treats response categories as approximately equally spaced, gives likelihood and impact equal weight, and allows partial compensation between them. LI-RII is interpreted as a descriptive within-study priority index. Factors are ranked in descending order, with equal source mean scores assigned average ranks.

The LI-RII values are obtained by dividing the reported two-decimal mean influence scores by 25 and are presented to three decimal places. Supplementary Tables S5, S6 provide the mean scores, normalized indices, and ranks for all 24 factors. Rankings use the source means rather than rounded display values. NPI scores are reported to four decimal places; equal reported NPI scores receive average ranks under the same tie rule. The Supplementary material provides the aggregate inputs supporting the rank comparisons, structural-reach analysis, and edge-deletion checks.

Step 2: Influence-state classification and Bayesian network modeling. The same respondent-level likelihood-impact products were converted into low, medium, and high states before network estimation. Figure 1 shows the classification matrix used for this purpose.

The operational classification bands were:

• High influence (red): score ≥ 15.
• Medium influence (yellow): score between 8 and 12.
• Low influence (green): score ≤ 6.

Each challenge and enabler produced an integer likelihood-impact product from 1 to 25. The attainable products are 1, 2, 3, 4, 5, 6, 8, 9, 10, 12, 15, 16, 20, and 25. The operational bands follow the cell colors in Figure 1: low (1–6), medium (8–12), and high (15–25). The only gaps between the bands are 7, 13, and 14, and none can arise from multiplying two integer ratings from 1 to 5. Every attainable product is therefore assigned to one state. This discretization preserves the influence-matrix logic while converting the variables into the three states required by the BBN. The same 15 challenges and nine enablers were retained across the LI-RII and BBN analyses.

A BBN is a probabilistic graphical model in which nodes represent variables and finalized directed edges encode modeled conditional dependencies (Khan et al., 2025; Tsagris, 2019). In this study, the BBN complemented LI-RII by examining how the same factors were positioned within a wider dependency structure. Its purpose was to explore modeled network propagation and the connected positions of the factors within that structure. The challenge and enabler network structures are shown in Figures 2, 3, respectively, while the corresponding NPI scores and BBN ranks are reported in Tables 5, 6.

Table 5. BBN NPI results for challenges.

NPI is a model-derived high-versus-low network-exposure contrast, reported to four decimal places. Equal reported scores receive average ranks: 5 for the three factors occupying positions 4 to 6 and 7.5 for the two occupying positions 7 to 8. Display order within a tied group does not imply a priority difference.

Table 6. BBN NPI results for enablers.

NPI is a model-derived high-versus-low network-exposure contrast, reported to four decimal places. Ranks are assigned in descending score order using the same average-rank convention as LI-RII. The reported enabler NPI scores contain no ties.

The BBN workflow comprised three analysis stages. First, respondent-level influence scores were discretized into the low, medium, and high states defined in Figure 1. Second, network structure was learned in February 2025 using the Peter-Clark (PC) constraint-based algorithm in Bayes Server Desktop 11.0 (Bayes Server Ltd.); Bayes Server 11.0 was released on January 7, 2025 (Bayes Server Ltd., 2026a). The analysis retained the software’s default PC settings. Bayes Server implements PC through marginal and conditional independence tests, with the significance level and maximum number of conditioning variables among the principal structure-learning controls (Bayes Server Ltd., 2026b). Accordingly, the significance threshold and maximum-conditioning depth were retained at their Bayes Server 11.0 defaults rather than user-specified values. The models were static, so temporal-order controls were not applicable. Edge orientations were those returned by the PC learning routine under the retained defaults, and no user-specified orientation or link constraints were applied. Third, the PC-derived network structures were passed directly to Bayes Server Parameter Learning to estimate the node probability distributions. Bayes Server uses the expectation–maximization (EM) algorithm for maximum-likelihood parameter estimation (Bayes Server Ltd., 2026c). With the software defaults retained, all node distributions were learned; Bayes Server’s documented default parameter-learning configuration uses log-likelihood convergence, uniform priors, and clustering initialization, while the remaining initialization, iteration/tolerance, concurrency, seed, and inference controls were also left at their version-default values (Bayes Server Ltd., 2026a; Bayes Server Ltd., 2026c). The complete final challenge and enabler adjacency lists are reported in the manuscript and Supplementary material. The final arrows are interpreted as PC-derived conditional dependencies consistent with the exploratory purpose of the analysis.

After structure and parameter estimation, high and low evidence states were imposed on each focal node and the corresponding probabilities were propagated through the network. Following the network-exposure formulation used in network-based risk prioritization (Qazi and Dikmen, 2021; Mismar et al., 2022), Equation 2 defines mean high-state exposure under evidence e, and Equation 3 defines the NPI contrast for node j.

RE ( e ) = 1 m ∑ k = 1 m P ( X k = High ∣ e ) (2)

NPI j = RE ( X j = High ) − RE ( X j = Low ) (3)

Here, Xk denotes an influence-state variable and m is the number of nodes included in the exposure average. A higher NPI indicates a larger difference in mean high-state exposure between the focal node’s High and Low evidence states. This probability-based contrast is distinct from the number of downstream nodes and from an estimated effect on observed NZEB adoption.

Method integration used factor-aligned rank comparison. Spearman’s rho and Kendall’s tau-b quantify correspondence between the standalone and NPI-based orderings, with average ranks used for ties in both. Sensitivity to distinctions within equal reported NPI scores was examined across all possible strict orderings within the tied groups. The intervention screen reports individual NPI contrasts alongside structural reach, which counts targeted enablers and their downstream descendants. For intervention set S, structural reach is defined in Equation 4.

R ( S ) = ∣ S ∪ Desc ( S ) ∣ ∣ V ∣ × 100 % (4)

In Equation 4, Desc(S) is the union of all downstream descendants of the targeted nodes, and V is the nine-node enabler set. The coordinated portfolio comprises Public Awareness Campaigns, Industry Training Programs, Government Incentives, and Better Financing. Reach was recalculated after removing each enabler edge in turn. An awareness-and-incentives pair provides a parsimonious structural comparator for assessing the contribution of directly targeting training and financing. Individual NPI values remain separate from collective structural reach. The integrated interpretation is organized into four intervention domains:

• Policy and Regulation encompasses codes, standards, and approval processes.
• Finance and Incentives encompasses financing access and targeted financial support.
• Capability and Integration encompasses workforce skills, technical coordination, and delivery capability.
• Public Engagement and Operation encompasses awareness, informed building use, and post-occupancy support.

These four domains organize the practical interpretation of the combined evidence. The analyses retain the same factor set, distinguish standalone priority from probabilistic propagation and structural coverage, and interpret the purposive sample as exploratory professional evidence.

Figure 4 summarizes the sequential exploratory mixed-methods design, showing the progression from literature-informed factor development to qualitative refinement, quantitative prioritization and network modeling, and integrated interpretation.

4 Results

The results connect qualitative factor development with two quantitative perspectives on priority. Section 4.1 presents the qualitative findings, Section 4.2 reports the standalone and network-based results, and Sections 4.3 and 4.4 examine their correspondence and the structural implications for coordinated intervention.

4.1 Qualitative analysis

Five semi-structured interviews generated the context-specific evidence used to refine the factor set for RQ1. The participants were mid-to-senior professionals with direct experience in sustainability, energy, engineering, consulting, or research. Table 7 summarizes their organizational settings, roles, and experience.

Table 7. Profiles of semi-structured interview participants.

The interviews covered regulatory development, technical integration, financing, new construction and retrofits, existing initiatives such as Shams Dubai, measurable performance targets, and collaboration among government, industry, utilities, and academia. The breadth of these topics reflects an active UAE policy, market, and technology environment and enabled the analysis to capture technical, institutional, economic, and behavioral dimensions of NZEB adoption rather than treating adoption as a single-technology decision.

The qualitative findings are presented at two analytical levels. Six descriptive coding categories summarize participant coverage, while five interpretive themes organize the substantive adoption issues. For interpretation, Policy and Regulatory Gaps are considered together within Policy and Regulatory Landscape; Economic and Social relate to Economic and Financial Conditions and Social Awareness and Market Engagement, respectively. Environmental and Resource considerations inform Technological Innovation and Infrastructure, while Retrofitting Opportunities is a cross-cutting theme concerning technical, resource, and financial conditions in existing buildings. The five themes therefore synthesize the reported findings rather than reproduce the six-category incidence matrix column by column.

• Policy and Regulatory Landscape: Participants recognized strong and growing national momentum through the UAE Net Zero by 2050 Strategic Initiative, Estidama, and Shams Dubai. They viewed continued harmonization of NZEB-specific requirements, performance expectations, and project implementation as a natural next step in this positive trajectory. Shams Dubai was repeatedly viewed as an important enabling foundation for wider market adoption. Participants therefore characterized the policy landscape as an established platform on which increasingly coordinated NZEB delivery can be built.
• Technological Innovation and Infrastructure: Participants highlighted major renewable-energy investments, including the Mohammed bin Rashid Al Maktoum Solar Park, as a strong platform for NZEB development. BIPV systems were viewed as promising for both energy generation and envelope performance, particularly in dense urban settings. Affordability, system integration, retrofit compatibility, and dependable operation were identified as areas where the UAE’s existing technology and infrastructure momentum can support broader project-level deployment.
• Economic and Financial Conditions: Participants emphasized opportunities to further strengthen the investment case for NZEBs, particularly for existing buildings. They recognized mechanisms such as low-interest loans and subsidies and identified accessible green financing, transparent life-cycle benefits, and targeted incentives as practical means of accelerating adoption. They viewed these mechanisms as natural extensions of the UAE’s broader sustainability and investment environment.
• Social Awareness and Market Engagement: Participants viewed public understanding and market visibility as important channels for strengthening demand and participation. They highlighted public education, demonstration projects, media visibility, and the integration of NZEB principles into academic and professional training as mechanisms through which the UAE can further extend engagement with high-performance and net-zero building practices.
• Retrofitting Opportunities: Participants viewed the existing building stock as an important opportunity for tailored NZEB solutions that account for spatial, structural, and system-compatibility conditions. Their responses favored incorporating NZEB readiness during initial design while developing building-specific retrofit strategies that can progressively extend high-performance practices across the existing stock.

Figure 5 presents participant coverage across the six descriptive coding categories. A filled circle denotes a recorded participant-category link. The matrix describes coverage rather than the frequency or importance of a category; Supplementary Table S4 provides the corresponding entries in participant order P1 to P5.

P3 was linked to all six categories, covering environmental, economic, resource, policy, regulatory, and social considerations. P1 and P5 were linked to the environmental, economic, and resource categories; P2 to the social category; and P4 to the policy and regulatory categories. These patterns describe the coverage recorded in the matrix and are interpreted alongside the five thematic accounts above.

Together, the themes connect policy and regulation, technology and infrastructure, finance, social engagement, and retrofitting. Table 8 organizes the resulting survey factors into practical implementation-factor and enabler groupings: integration, costs and capability, regulation, behavior, financing, awareness, and supporting technologies. These factor groupings link the thematic interpretation to the 15 challenge factors and nine enablers assessed quantitatively.

Table 8. Qualitative findings and NZEB survey-factor groupings in the UAE.

The factor groupings in Table 8 serve as the bridge to quantitative analysis. BIPV and stronger institutional engagement remain contextual findings, while the 15 challenges and nine enablers form the common LI-RII and BBN factor set.

4.2 Quantitative analysis

The survey operationalized the factor set in Table 8 by obtaining separate likelihood and impact ratings for each challenge and enabler from professionals involved in construction, sustainability, energy, design, or regulation in the UAE.

Thirty responses were analyzed. Engineers represented 73.3% of the sample and sustainability experts 13.3%; the remaining respondents represented policymaking, architecture, and design management.

Because the sample was purposive, small, and engineer-dominant, the quantitative findings are reported as exploratory professional judgments rather than population estimates. Their analytical value lies in comparing standalone priorities and modeled network positions within the same defined factor set.

Figure 6 summarizes the professional composition of the survey sample. Engineers constitute 73.3% of respondents, followed by sustainability experts at 13.3%, with the remaining respondents representing policymaking, architecture, and design management.

The predominance of engineers is relevant because many survey factors concern building systems, technical integration, delivery capability, and energy-performance management.

The smaller professional groups contribute sustainability, architectural, regulatory, and project-coordination perspectives, but the quantitative findings should be interpreted primarily as technically informed professional assessments.

The quantitative analysis first reports the LI-RII ordering before considering modeled interdependencies. The BBN/NPI results are then presented separately, followed by factor-aligned comparison and robustness analysis.

4.2.1 LI-RII-based independent ranking analysis

Tables 9, 10 report the challenge and enabler priorities for RQ1. LI-RII values express the normalized mean likelihood-impact product and are reported to three decimal places; ranks retain distinctions in the source means and use average ranks for ties.

Table 9. LI-RII results for challenges.

LI-RII equals the reported two-decimal mean influence score divided by 25 and is displayed to three decimal places. Ranks use source means; equal means receive average ranks. Higher values indicate greater standalone priority.

Table 10. LI-RII results for enablers.

LI-RII equals the reported two-decimal mean influence score divided by 25 and is displayed to three decimal places. Ranks use source means; equal means receive average ranks. Higher values indicate greater standalone priority.

4.2.1.1 Challenge LI-RII results

Occupant-behavior considerations rank first among the challenge factors (LI-RII = 0.483), followed by approval-process duration (0.421) and workforce capacity (0.416). Specialist labor cost ranks fourth (0.403), and opportunities to strengthen public awareness rank fifth (0.388). Table 9 reports the complete ordering using the exact survey-factor labels.

The challenge-factor ordering concentrates standalone priority in building operation, approval efficiency, workforce capacity and cost, public engagement, and supply-chain readiness. These priorities identify areas where continued coordination can support efficient NZEB implementation.

4.2.1.2 Enabler LI-RII results

Public Awareness Campaigns rank first among the enablers (LI-RII = 0.409), followed by Renewable Energy Storage Technologies (0.385), Better Financing (0.377), and Industry Training Programs (0.372). Table 10 reports the complete ordering.

The enabler ordering gives greatest standalone priority to public awareness, followed by technology availability, financing, and industry capability. It therefore describes what respondents would prioritize when each enabling action is considered separately.

The LI-RII results therefore answer the independent-priority component of RQ1. They do not indicate whether a factor is upstream, downstream, or widely connected within the adoption system. The BBN analysis addresses that separate question.

4.2.2 Interdependence-based analysis: BBN

The BBN used the same 15 challenges and nine enablers as the LI-RII analysis. The challenge and enabler networks are presented separately, followed by their NPI scores and ranks. Network directions are interpreted as modeled conditional dependencies consistent with the exploratory purpose of the analysis. The LI-RII and BBN results are integrated in Section 4.3.

4.2.2.1 Challenge BBN results

Figure 2 presents the final challenge-factor Bayesian belief network generated in Bayes Server Desktop 11.0. Each node displays the modeled distribution across the high, medium, and low influence states, while directed edges represent the PC-derived conditional dependencies in the structure used for parameter learning. The topology provides the structural basis for the NPI results reported in Table 5.

The PC algorithm generated the network connections and orientations used directly for subsequent parameter learning. Table 11 reports the complete final challenge adjacency and provides a cautious substantive interpretation of every learned link. These interpretations are used to examine implementation coherence while the network directions themselves are treated as exploratory conditional dependencies rather than confirmed causal effects.

Table 11. Modeled challenge links and substantive interpretations.

The final challenge topology contains several connected pathways. System-integration requirements, opportunities to strengthen public awareness, and supply-chain readiness connect to initial capital requirements. Initial capital requirements then connect to access to financial incentives, which connects to developer engagement and subsequently to occupant-behavior considerations. Supply-chain readiness also connects to zoning alignment, while zoning alignment and building-code clarity both connect to approval-process duration. Workforce capacity connects to specialist labor cost and compatibility requirements.

These links describe the finalized exploratory model and provide testable hypotheses about implementation dependencies. Their value lies in identifying connected leverage points that can be examined in larger, longitudinal, or project-level studies.

Table 5 reports the challenge NPI scores. System-integration requirements, initial capital requirements, and occupant-behavior considerations occupy the first three positions. Approval-process duration, building-code clarity, and access to financial incentives share NPI = 0.0667 and average rank 5. Workforce capacity and compatibility requirements share NPI = 0.0594 and average rank 7.5. The table retains the exact factor labels used in the model.

The challenge NPI ordering highlights integration, capital requirements, and building operation as prominent components of modeled network exposure.

The same analytical sequence is then applied to the enablers. Figure 3 presents the enabler network, and Table 6 reports the corresponding NPI scores and BBN ranks.

4.2.2.2 Enabler BBN results

Figure 3 presents the final enabler-factor Bayesian belief network generated in Bayes Server Desktop 11.0. Each node displays the modeled distribution across the high, medium, and low influence states, while directed edges represent the PC-derived conditional dependencies among the nine enablers in the structure used for parameter learning. The topology supports the NPI ranking in Table 6 and the structural-reach analysis in Section 4.4.

The final topology links Government Incentives to Better Financing and Fast-Track Permitting. Better Financing links to Stricter Building Codes and Renewable Energy Storage Technologies, while Stricter Building Codes also links to Fast-Track Permitting. These directions show that the policy and financing nodes were modeled as connected rather than isolated measures.

A second chain links Public Awareness Campaigns to Industry Training Programs and then to Smart Energy Management. Within the exploratory model, this sequence is consistent with an interpretation in which awareness supports capability development and training supports advanced-system implementation, but the arrows do not establish temporal or causal order.

Circular Economy Principles and Better Financing both link to Renewable Energy Storage Technologies. The downstream position of storage in the final graph does not diminish its standalone importance; it indicates only its modeled position within the exploratory network.

Table 6 reports the complete enabler NPI results. Public Awareness Campaigns rank first, followed by Industry Training Programs and Government Incentives. Stricter Building Codes and Fast-Track Permitting occupy the fourth and fifth positions. The ranking therefore emphasizes public engagement, workforce capability, policy support, regulatory clarity, and permitting as the enablers with the highest NPI values.

The enabler BBN results are therefore interpreted as a network-level ordering of enabling conditions. Their relationship to standalone LI-RII priority is examined in the following comparison.

4.3 Comparison of LI-RII and BBN rankings

Tables 12, 13 compare factor-aligned ranks to answer RQ2. LI-RII represents standalone stakeholder priority, whereas NPI represents a modeled network-exposure contrast. Comparing ranks places these distinct constructs alongside one another without combining their numerical scales.

Table 12. Comparison of LI-RII and BBN ranks for challenges.

Both columns use average ranks for ties. Rank change equals LI-RII rank minus BBN rank; a positive change indicates a higher relative position under NPI. Compatibility Issues and Resistance from Developers share LI-RII rank 14.5; the NPI tied groups receive ranks 5 and 7.5.

Table 13. Comparison of LI-RII and BBN ranks for enablers.

“Up” means that a factor occupies a stronger priority position in the BBN than under LI-RII. The comparison uses ranks only because LI-RII and NPI are different metrics.

4.3.1 Challenge rank comparison

The challenge-factor rankings differ materially. LI-RII places occupant-behavior considerations, approval-process duration, and workforce capacity in the first three positions, whereas the BBN places system-integration requirements, initial capital requirements, and occupant-behavior considerations first. Occupant-behavior considerations therefore appear in the top three under both methods.

The largest upward movements are system-integration requirements, from rank 12 under LI-RII to rank 1 under NPI, and initial capital requirements, from rank 13 to rank 2. Access to financial incentives rises from rank 11 to tied average rank 5. Specialist labor cost moves from rank 4 to rank 11, supply-chain readiness from rank 6 to rank 12, and workforce capacity from rank 3 to tied average rank 7.5. Table 12 reports the full comparison using average ranks consistently in both columns.

The divergence is decision-relevant rather than contradictory. LI-RII identifies implementation considerations that respondents view as most immediate when assessed independently, whereas NPI identifies factors whose modeled state changes produce larger contrasts in network exposure. The two methods therefore answer complementary questions.

The comparison shows that considering interdependencies changes the factor ordering substantially. Immediate implementation priorities center on occupant behavior, approval efficiency, and workforce capacity; modeled network leverage centers on integration, capital requirements, and occupant behavior.

4.3.2 Enabler rank comparison

The enabler rankings show a different pattern. Public Awareness Campaigns remain first under both methods, but Industry Training Programs rise from fourth under LI-RII to second in the BBN, and Government Incentives rise from ninth to third.

These upward movements are consistent with the fitted network connections of training and incentives to several other enabling conditions. Training is linked with implementation capability and advanced-system operation, while incentives are linked with financing, permitting, and regulatory support.

Renewable Energy Storage Technologies moves from rank 2 under LI-RII to rank 9 under NPI, and Better Financing moves from rank 3 to rank 8. Their high standalone priority remains important to the integrated interpretation. Their lower NPI positions indicate smaller modeled exposure contrasts rather than lower practical importance or fewer downstream connections.

Table 13 summarizes the full rank comparison. Public Awareness Campaigns are the clearest cross-method enabler, while Industry Training Programs and Government Incentives emerge as the strongest network-based complements to the independently prioritized technology and financing measures.

Taken together, the results answer RQ1 and RQ2 at an exploratory level. The qualitative phase identifies the UAE-specific factor set; LI-RII establishes standalone priorities; BBN/NPI establishes modeled network propagation; and the factor-aligned comparison shows low rank correspondence for both challenges and enablers. These differences provide the basis for the coordinated pathway developed in Section 5.

4.4 Rank-correlation, scenario, and robustness checks

Spearman’s rho and Kendall’s tau-b were calculated from the factor-aligned average ranks in Tables 12, 13. Challenge-rank correspondence is weak (rho = 0.0673, p = 0.8116; tau-b = 0.0976, p = 0.6183). Enabler-rank correspondence is negligible (rho = −0.0167, p = 0.9661; tau-b = −0.0556, p = 0.9195). The coefficients describe correspondence across the study’s fixed factor sets; the p values are supporting diagnostics rather than evidence of population-level agreement or independence. Across all 12 strict orderings within the two tied NPI groups, challenge rho ranges from 0.0143 to 0.1197 and tau-b from 0.0574 to 0.1340. Thus, the weak-correspondence interpretation is retained under every such ordering (Supplementary Table S10).

The intervention screen reports individual NPI contrasts alongside structural reach in the nine-node enabler network. Government Incentives has NPI = 0.1326 and reaches five nodes (55.6%); Better Financing has NPI = 0.0917 and reaches four (44.4%); Public Awareness Campaigns has NPI = 0.1440 and reaches three (33.3%); and Industry Training Programs has NPI = 0.1333 and reaches two (22.2%). The coordinated four-action portfolio reaches eight nodes (88.9%) and retains seven or eight (77.8 to 88.9%) after any single enabler edge is removed. Table 14 summarizes the individual and coordinated scenarios. These results quantify coverage of enabling conditions within the modeled topology.

Table 14. Intervention-oriented scenario screen for enablers.

Individual NPI values represent separate high-versus-low network-exposure contrasts for each focal enabler. Because these effects overlap within the network, individual NPI values are not additive. The coordinated portfolio is therefore evaluated using its combined structural reach rather than the sum of individual NPI values.

Awareness and incentives alone also reach eight nodes in the intact topology. Their coverage ranges from five to eight nodes under the same edge-deletion checks, compared with seven to eight for the four-action portfolio (Supplementary Table S11). Directly including training and financing therefore adds structural redundancy rather than increasing intact-network coverage.

5 Discussion

The combined findings answer RQ1 and RQ2 through a clear pattern of complementary divergence. LI-RII emphasizes the implementation challenge factors and enablers judged most important when considered independently, while BBN/NPI emphasizes nodes with broader modeled propagation. For challenge factors, the low rank correspondence (rho = 0.0673; tau-b = 0.0976) distinguishes immediate implementation priorities from network leverage. For enablers, Public Awareness Campaigns remain first under both methods, while the near-zero overall correspondence (rho = −0.0167; tau-b = −0.0556) shows that training and incentives become more prominent once interdependencies are considered. Both results arise from the same exploratory survey and provide complementary perspectives within an integrated decision-support analysis. Taken together, the findings identify where additional coordination can generate value within the UAE’s established sustainability and net-zero ecosystem and complement the country’s existing policy, green-building, renewable-energy, and industry-capability initiatives (Awadh, 2018; Dubai Electricity and Water Authority, 2024; Emirates Green Building Council, 2024; Asfour and Mansour, 2026; Pandita et al., 2024; Fayyad and John, 2017; Mistarihi et al., 2025).

The challenge-factor comparison separates immediate implementation priorities from network-based exposure contrasts. Occupant-behavior considerations, approval-process duration, and workforce capacity lead the standalone ordering. System-integration requirements and initial capital requirements become more prominent under NPI. In the challenge topology, integration connects to capital requirements, which connect through access to financial support and developer engagement to occupant-behavior considerations. Workforce capability separately connects to specialist labor cost and compatibility requirements, while building-code clarity and zoning alignment connect to approval efficiency. These relationships identify opportunities to reinforce the UAE’s established sustainability progress by coordinating delivery efficiency, financial support, and technical capability within an already active implementation environment.

The prominence of occupant-behavior considerations adds an essential operational dimension. This factor ranks first under LI-RII and third under BBN/NPI, but it should not be interpreted as assigning responsibility for performance outcomes to occupants. High-performance buildings realize their designed potential through the continuing interaction of users, controls, facility management, maintenance, commissioning, and operational information. In hot-arid settings such as the UAE, where efficient cooling is central to building performance, user-centered controls and effective operational practices can materially strengthen realized performance. Recent performance-gap research similarly locates occupant-related outcomes within broader interactions among design assumptions, control systems, management practices, and lifecycle information (Zheng et al., 2024; Mahdavi and Berger, 2025).

The enabler comparison adds a different form of insight. Public Awareness Campaigns remain the clearest cross-method priority. Renewable Energy Storage Technologies and Better Financing rank highly under LI-RII, while Industry Training Programs and Government Incentives rise under BBN/NPI. The finalized topology links awareness to training and training to smart energy management; it also links incentives to financing and permitting, and financing to codes and storage. The scenario screen makes that distinction actionable: awareness has the largest individual NPI, incentives have the widest single-action structural reach, and the four-action portfolio covers eight of nine enabler nodes. Storage and financing therefore remain important, but their system contribution is strengthened when awareness, capability, and policy support are activated together. This combination aligns closely with the UAE’s existing policy and institutional instruments, which already connect sustainability standards, distributed solar participation, and industry engagement (Awadh, 2018; Dubai Electricity and Water Authority, 2024; Emirates Green Building Council, 2024; Asfour and Mansour, 2026).

RQ3 is addressed through a coordinated portfolio combining awareness, training, incentives, and financing. The enabler topology links awareness with training and smart energy management, and links incentives and financing with codes, permitting, and storage. Direct support for all four actions provides coverage across these pathways and structural redundancy under individual edge deletions. The practical interpretation spans Policy and Regulation, Finance and Incentives, Capability and Integration, and Public Engagement and Operation. Figure 7 integrates the standalone priorities, NPI-based rankings, and these four intervention domains. The portfolio therefore functions as a coordination layer for amplifying existing UAE initiatives and aligning their effects across policy, finance, capability, and public engagement.

6 Conclusions and future research

Against the backdrop of the UAE’s strong national sustainability agenda, this study examined how standalone stakeholder priorities and modeled interdependencies jointly shape NZEB adoption decisions. The LI-RII analysis highlights occupant-behavior considerations, approval-process duration, and workforce capacity as the leading standalone challenge priorities, whereas the BBN/NPI analysis highlights system-integration requirements, initial capital requirements, and occupant-behavior considerations as the leading network positions. Challenge-rank correspondence is low. Public Awareness Campaigns rank first among enablers under both methods, while Industry Training Programs and Government Incentives become more prominent in the network model. The central contribution is the explicit distinction between immediate stakeholder priority and modeled network leverage, providing a complementary perspective for further coordinating and amplifying the UAE’s established sustainability, green-building, renewable-energy, and net-zero implementation foundations.

RQ3 is answered through a coordinated four-action portfolio comprising public awareness, industry training, government incentives, and better financing. The final topology shows that this portfolio reaches eight of nine enabler nodes, compared with five for incentives, four for financing, three for awareness, and two for training individually. Leave-one-edge-out analysis retains seven to eight nodes of combined coverage. Continued code alignment, efficient approval pathways, workforce development, technical integration, and post-occupancy support can therefore be advanced as mutually reinforcing implementation components that build on the UAE’s strong policy, market, and technology foundations. The portfolio is positioned as a means of connecting and amplifying existing UAE strengths across policy, finance, workforce capability, technology, and public engagement. Taken together, the findings reinforce the UAE’s strong position for continued NZEB advancement, supported by national net-zero strategy, established green-building systems, distributed-solar initiatives, renewable-energy investment, and growing technical capability (Awadh, 2018; Dubai Electricity and Water Authority, 2024; Emirates Green Building Council, 2024; Asfour and Mansour, 2026; Pandita et al., 2024; Fayyad and John, 2017; Mistarihi et al., 2025).

The findings should be interpreted within the study’s exploratory boundaries. Five information-rich interviews supported context-specific factor refinement, and 30 survey responses supported descriptive prioritization and two separate exploratory networks of 15 and nine nodes. The strong engineering representation in the survey (73.3%) provides a technically informed perspective, giving particular salience to system integration, workforce capability, commissioning, and technical operation while complementary professional perspectives remain relevant to future validation. The BBN arrows and NPI values represent modeled conditional dependencies and network-exposure contrasts within this exploratory design. Transferability beyond the UAE is framed as a set of testable propositions for other hot-arid settings, providing a clear basis for external validation.

Analytical transparency is supported through disclosure of the aggregate factor scores and ranks, complete final adjacency structures, NPI values, factor-aligned rank calculations, structural-reach scenarios, and leave-one-edge-out results in the Supplementary material. This reporting enables reproduction of the analyses at the aggregate and final-topology levels while retaining the study’s exploratory scope.

Future research should examine the proposed dependencies with larger, more professionally diverse samples across emirates, including planners, facility managers, developers, financiers, building operators, and utilities. Repeated structure learning and sensitivity analysis of discretization thresholds and conditional probabilities can assess model stability. Extending the network with an observed NZEB-adoption or implementation-readiness outcome would support posterior comparisons of individual and joint intervention scenarios. Longitudinal project studies and comparative work across Gulf Cooperation Council and other hot-arid regions can assess which relationships transfer across settings, while life-cycle costing can test the resource implications of alternative portfolios. Comparative testing of alternative aggregation rules can also assess the sensitivity of standalone priorities to the chosen index.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frsus.2026.1959319/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Long Zhang, Tianjin University of Technology, China

Reviewed by: Syed Shahid Shah, Beihang University, China

Fuat Emre Kaya, University of Sassari, Italy