ARISE: an urban political economy-informed resilience indicator framework for south asian metropolitan cities

1 Introduction

Cities act as dynamic engines of economic development and technological innovation while also serving as epicentres of socio-environmental risk. In data-constrained urban contexts, the rapid change from primary economy to urban economies has intensified both development pressures and vulnerabilities of urban systems. Global disruptions, such as the 2008 financial recession and the 2020 COVID-19 pandemic, have greatly increased the imperative to capacitate cities that can absorb, recover from, and transform as a response to sudden dynamic shocks (Brada et al., 2021; Fang et al., 2025).

Urban resilience, defined as the capacity of urban systems to absorb, recover from, and adapt to sudden disruptions while sustaining core functions, has become a foundational concept in planning, governance, and development scholarship (Meerow and Newell, 2019; Sharifi, 2023; Kapucu et al., 2024). Yet existing resilience assessment frameworks remain inadequate for data-constrained urban contexts, insufficiently integrating political-economic context, stakeholder participation, and governance aspects which are critical for urban resilience assessment (Etinay et al., 2018; Tyler et al., 2016).

This study introduces ARISE, the Adaptive Resilience Indicator System for Equity, a systematic five-stage Urban Political Economy (UPE)-informed framework for identifying, categorising, reducing, validating, and prioritising 32 Resilience Indicators for urban resilience assessment. The methodological pipeline this framework employs, literature-derived pooling, frequency-based reduction, expert validation, and multi-criteria weighting, is, on its own, a well-established approach already widely used in sustainability indicator research (Gani et al., 2021; 2022; Pei et al., 2019). This paper’s contribution does not rest on novelty at the level of these individual techniques. It rests on demonstrating, with specific and checkable examples at each pipeline stage, how a UPE theoretical commitment can be made to do substantive work throughout an indicator-development process, conditioning which categories of indicator are considered admissible in the first place, protecting theoretically motivated but empirically under-cited indicators from elimination by frequency alone, and, critically, providing the interpretive lens through which the framework’s own empirical results are read back against its theoretical premises rather than reported as self-evidently consistent with them. The framework addresses four research questions, the last of which is the paper’s central methodological question and is treated as such throughout. The need for developing such a framework is due to the magnitude of the challenge as suggested by the United Nations projections that by 2050, two-thirds of the global population would inevitably settle in cities, with most of the low-quality settlement establishments in Africa and Asia (United Nations, 2018). Cities in lower- and middle-income countries are simultaneously the primary sites of poverty concentration, ecological pressure, and governance fragmentation—precisely the settings where resilience frameworks are most needed but least available. Existing tools, calibrated to the institutional environments and data infrastructures of high-income cities, fail systematically when applied to these contexts. ARISE responds to this gap by building an indicator framework whose design logic begins from the realities of data-constrained South Asian metropolitan cities.

• Does the indicator set identify resource consumption patterns and urban synergies advocating regeneration and intragenerational equity?
• Does the indicator set identify trade-off patterns between economic, ecological, and social goals set for the urban systems?
• Does the indicator set assess the equitability of infrastructure and governance flows that strengthen urban resilience?
• [Central Question] What added value does UPE-informed thinking bring to resilience-based urban studies, and how cohesively does ARISE integrate the UPE and the essence of urban resilience relationship? This question structures the analysis throughout the paper and is revisited explicitly in Section 5.1.1 and the Conclusion.

2 Literature review

2.1 Urban resilience: conceptual evolution

Resilience as a concept entered urban scholarship from Holling’s (1973) ecological systems theory and has since bifurcated into an engineering tradition emphasising bounce-back speed, and a political-economic tradition foregrounding distributional questions of who has access to resilience resources and who bears disturbance costs (Meerow and Newell, 2019; Folke, 2006). The political-economic tradition, upon which ARISE is grounded, treats urban resilience not as a generic system property but as a contested outcome shaped by governance arrangements, resource allocation politics, and spatial inequality (Figure 1 illustrates this conceptual evolution).

Contemporary theories of urban resilience have established three perspectives: engineering resilience, foregrounding the process of return to the equilibrium state; ecological resilience, which emphasises maintaining persistence under disturbance; and evolutionary resilience, which emphasises adaptive transformation within non-linear trajectories (Sharifi, 2023). Meerow and Newell’s (2019) critical extension—asking resilience for whom, what, when, where, and why—foregrounds distributional and political dimensions that are foundational to this study. Resilience is not a politically neutral property; it is shaped by decisions about who receives infrastructure investment, whose knowledge is incorporated into governance processes, and whose vulnerability is prioritised in planning responses. The ARISE framework proceeds from political-economic processes rather than purely technical assessment. This framework advances established critical theory into an operational methodology. It integrates political economy dynamics natively within its quantitatively weighted metrics, moving beyond mere contextual observation.

2.2 Urban political economy–the theoretical foundation of ARISE

Urban Political Economy (UPE) emerged from critical geography and political ecology to examine the relationship between capital, power, and urban space (Harvey, 1989; Swyngedouw, 1996). UPE is centrally concerned with primary political decisions, market forces, and institutional arrangements produce uneven urban development, shaping who benefits from urban infrastructure, which bears the burden of the environment, and the knowledge of informed governance (Heynen et al., 2006). ARISE translates UPE’s core concepts directly into indicator design. Urban metabolism represents the flows of energy, water, materials, and nutrients through various urban systems (Kennedy et al., 2011), and operationalised through dimensions such as Ecosystem and Resources. Circular resource flows and material footprint indicators are increasingly used to operationalise urban metabolism at policy-relevant scales (Bahers and Rosado, 2023). Politicised Resource Flow reflects UPE’s insistence that resource allocation is never politically neutral. The Power Devolved to LPA reflects UPE’s notion of decentralisation of governance as a prerequisite for equity-responsive planning. ARISE is therefore not merely a technical indicator framework but a politically grounded assessment tool whose design criteria are theoretically derived from UPE principles. This theoretical grounding distinguishes ARISE from frameworks that name equity or governance as desirable properties while selecting indicators on empirical frequency alone. In ARISE, indicator selection is determined not merely by citation frequency but by asking which dimensions of urban life UPE theory identifies as constitutive of uneven development. This produces a diverse indicator set than other technocratic approaches that would generate: one where the political economy of resource flows, the spatial distribution of vulnerable areas, and the structure of governance setup are represented to be primary rather than secondary variables. The significance of these gaps becomes apparent in Table 1: every existing framework addresses governance only partially, and none designs for data-constrained urban applicability using UPE-derived selection criteria.

TABLE 1. Comparative gap analysis: ARISE versus existing resilience assessment frameworks.

Coding criteria—Yes: at least one named, defined indicator explicitly addressing the dimension; Partial: dimension mentioned in framework narrative but not operationalised as a distinct indicator; No: no substantive treatment. Hudec et al. (2018) correctly classified as Spatial = Yes (applied across 79 Slovak districts). Frameworks reviewed as of the versions cited. Source: Author’s synthesis.

2.3 Existing frameworks: a comparative gap analysis

A systematic comparison of six leading resilience assessment frameworks reveals persistent coverage gaps, particularly with respect to data-constrained urban contexts (Table 1).

Two additional frameworks in Table 1 sharpen the comparison: Krishnan et al. (2023) RISE-UP, the closest comparator to ARISE in the South Asian literature, covers spatial and data-constrained dimensions but classifies governance as partial without indicator-level disaggregation and lacks UPE-derived categorisation. The City Resilience Index (Resilient Cities Network, 2019), deployed in 100+ cities globally, achieves governance and social coverage but requires facilitation infrastructure limiting applicability in data-constrained settings. Khazai et al., (2018) Resilience Performance Scorecard offers expert-validated, spatially explicit assessment but limited governance disaggregation and no UPE grounding. ARISE addresses all six Table 1 dimensions, disaggregates governance into four distinct indicators with individual weights, and uniquely grounds its categorisation in UPE theory rather than empirical frequency alone.

2.4 Urban resilience indicators: methodological challenges

Effective resilience indicators must satisfy four criteria: representativeness (capturing full urban system scope); responsiveness (sensitivity to dynamic change); measurability (operationalisable through accessible data); and non-redundancy (each indicator providing unique informational content) (Etinay et al., 2018; Pandey and Ghosh, 2023; Urquiza et al., 2021). Existing indicator sets suffer from subjective selection without systematic reduction; single-scale assessments across heterogeneous contexts; absence of stakeholder participation in validation; and inadequate attention to data-constrained urban data environments (Saja et al., 2019; Sharifi, 2016). This study addresses each limitation through the five-stage methodology described in Section 3. The consequences of these methodological weaknesses are not merely academic. When indicator sets are assembled through ad hoc expert opinion without systematic reduction, the resulting frameworks tend to be unwieldy, with some published sets containing over 200 indicators, far exceeding the practical capacity of planning institutions to monitor and report against them (Sharifi, 2016). When geographical scope is restricted to high-income country contexts, the data requirements, institutional arrangements, and governance structures built into the framework become inapplicable in the contexts where resilience planning is most urgently needed. When stakeholder participation is absent from the validation stage, the indicator weights reflect academic preferences rather than the judgements of domain specialists who actually design, implement, and evaluate urban planning and infrastructure systems. And when data-constrained urban data environments are not accounted for, indicators that require continuous satellite monitoring, real-time sensor networks, or comprehensive cadastral databases cannot be operationalised in the cities for which they are theoretically intended. ARISE is designed from the outset to address all four of these limitations simultaneously, not sequentially.

3 Methodology: the ARISE framework

ARISE adopts a mixed-methods approach integrating systematic literature synthesis, Pareto-based indicator reduction, expert consultation, and the Analytic Hierarchy Process (AHP). The methodology proceeds through five sequential stages (Figure 2): (1) Identification, (2) Categorisation, (3) Reduction, (4) Validation, and (5) Prioritisation.

3.1 Stage 1: indicator identification

Following the PRISMA 2020 framework (Page et al., 2021), a systematic review across Web of Science, Scopus, and Google Scholar used Boolean search terms combining ‘urban resilience,’ ‘resilience indicators,’ ‘multi-criteria analysis,’ ‘adaptive governance,’ and related terms (2013–2025). Citation-frequency thresholds provide one recognised, if imperfect, proxy for research relevance (Aksnes et al., 2019). Screening: 847 records identified; 312 duplicates removed; 535 screened at title/abstract; 398 excluded; 137 full-text reviewed; 62 excluded; 75 included. Geographic distribution: 45.3% data-constrained urban contexts, 29.3% mixed/comparative, 25.4% high-income contexts. The full literature pool is in Supplementary File Sheet S1.

3.2 Stage 2: categorisation

Following established principles for indicator set construction in sustainability assessment, including representativeness, measurability, and non-redundancy across the full conceptual domain (Bell and Morse, 2008), the 193 indicators were organised into seven thematic dimensions, yielding 124 RIs after categorisation. Indicators eliminated at this stage failed one or more of four criteria: (a) lacked empirical backing leading to speculation of impact on urban resilience outcomes, (b) were operationally infeasible in data-constrained urban data environments, (c) were substantively redundant with a retained indicator, or (d) were completely irrelevant to characteristics of urban systems such as governance, equity, or planning. Retained indicators were evaluated against:

The categorisation and exclusion process that reduced the initial pool of 193 indicators to 124 applied four criteria sequentially to each candidate indicator, and an auditable record of every decision is provided in Supplementary File 2, Supplementary Table S1. The four criteria, derived from established sustainability indicator-quality standards (Bell and Morse, 2008), were: (a) Conceptual Relevance, defined as whether the indicator captures a distinct, UPE-grounded dimension of urban resilience that is not already represented by a higher-frequency indicator in the retained set; (b) Empirical Evidence, defined as whether the indicator is supported by at least two independent peer-reviewed applications in the 75-paper literature pool rather than appearing in a single study; (c) Non-Redundancy, defined as whether the indicator captures information that is not subsumed by a conceptually broader, higher-frequency indicator already carried forward; and (d) Operational Feasibility, defined as whether the indicator is measurable using data that can reasonably be expected to exist at the urban local body level in lower- and middle-income country contexts. Indicators failing any one of these four criteria were excluded.

Of the 193 indicators, 69 were excluded under one or more of these criteria, leaving 124 indicators across seven thematic dimensions. Supplementary Table S1 (Supplementary File 1) documents every excluded indicator by name, paper-frequency count, and the primary criterion under which it was excluded. The dominant reason for exclusion was low empirical evidence combined with conceptual redundancy: 47 of the 69 excluded indicators (68.1%) had a frequency count of two or fewer across the 75 studies, meaning they appeared in at most two independent papers and therefore lacked the minimum replication required to establish cross-contextual relevance. Fourteen indicators (20.3%) were excluded under the non-redundancy criterion, having been subsumed by a higher-frequency indicator already retained in the same conceptual category; for example, Flood Resilience (n = 2) was subsumed by the retained Resilient Infrastructure (n = 12), and Traffic Volume Modeling (n = 1) was subsumed by the retained Accessibility (n = 11). Eight indicators (11.6%) were excluded primarily under the operational feasibility criterion as indicators whose measurement requires data not routinely produced by urban local bodies in data-constrained urban contexts, such as Biocapacity (n = 3), Resilience Skewness (n = 4), and Global Infrastructure Index (n = 1).

The 124 retained indicators distributed across seven dimensions: Urban Design and Spatial Planning (19), Disaster Preparedness (15), Ecosystem and Resources (17), Economic Environment (16), Infrastructure and Logistics (23), Health and Wellbeing (25), and Governance (9). Governance’s smallest retained pool is consistent with UPE’s observation that governance dimensions are structurally under-studied relative to their theoretical importance (Meerow and Newell, 2019), motivating the two-track retention mechanism in Stage 3.

The exclusion criteria were applied consistently: Conceptual Relevance excluded indicators from incompatible theoretical traditions (systems-ecology, metabolic frameworks); Empirical Evidence excluded single-study indicators appearing in specialised data contexts not available in data-constrained urban settings; Non-Redundancy followed the double-counting avoidance practice of Kutty et al. (2022) and Sharifi (2016); and Operational Feasibility excluded indicators requiring data not routinely produced by urban local bodies in rapidly urbanising cities, including Biocapacity (Krishnan et al., 2023), Resilience Skewness, and Global Infrastructure Index.

• Conceptual Relevance: Indicators directly impacting urban resilience, including institutional capacity, vulnerability disparities, and governance quality, were retained.
• Empirical Evidence: Each indicator required a demonstrated effect on resilience outcomes in peer-reviewed research.
• Consolidation and Non-Redundancy: Overlapping indicators were merged into composite concepts to eliminate duplication.
• Operational Feasibility: Indicators required measurability and trackability within real planning and institutional contexts. The application of these four criteria to the 193-indicator initial pool resulted in the elimination of 69 indicators at Stage 2, representing a 35.8% reduction before any frequency-based filtering was applied. The dominant reason for exclusion was low empirical evidence combined with conceptual redundancy: 47 of the 69 excluded indicators (68.1%) had a frequency count of two or fewer across the 75 studies, meaning they appeared in at most two independent papers and therefore lacked the minimum cross-contextual replication required to establish generalised relevance. This includes indicators such as Transit Accessibility (n = 1), Traffic Volume Modeling (n = 1), and Participatory Mapping Technology (n = 2). Fourteen indicators (20.3%) were excluded under the non-redundancy criterion, having been subsumed by a higher-frequency indicator in the same conceptual category: for example, Flood Resilience (n = 2) was subsumed by the retained Resilient Infrastructure (n = 12), Land Use Mix Index was subsumed by Heterogeneous Resilient Planning, and Sustainable Supply Chain was subsumed by Supply Network. Eight indicators (11.6%) were excluded under the operational feasibility criterion because their measurement requires data not routinely produced by urban local bodies in data-constrained urban contexts: Biocapacity (n = 3) requires city-level ecological footprint data unavailable in India or most of data-constrained urban contexts in South Asia and comparable settings; Resilience Skewness (n = 4) is a statistical derivative of the framework’s outputs rather than an independent input; and Global Infrastructure Index (n = 1) requires national-scale data not computable at city level. The complete indicator-by-indicator exclusion record, including all 69 excluded indicators with dual criteria codes and specific reasons, is provided in Supplementary Table S1.

3.3 Stage 3: indicator reduction via pareto analysis

Stage 3 employs Pareto analysis to identify indicators contributing to 80% of cumulative citation frequency. Indicators were ranked by appearance frequency; those in the top 80% are classified as Vital (37 indicators) and those below as Useful (87 indicators). The threshold is anchored to Juran (1951)’s quality-management convention.

To assess the sensitivity of the final indicator set to the 80% threshold choice, the analysis was repeated at thresholds of 70% and 90%. At the 70% threshold, 21 indicators (rather than 37) are classified as Vital, with Community Participation, Accessibility, and Politicised Resource Flow dropping from the Vital to the Useful tier. At the 90% threshold, 53 indicators are classified as Vital, and the expert panel’s Stage 4 reduction would have had proportionally more material to evaluate. The 80% threshold produced a Vital tier of 37 indicators that the expert panel could review within a single SGT session of reasonable duration; both alternative thresholds would have either under-represented equity and governance indicators (at 70%) or generated a validation burden impractical for a panel of 18 specialists (at 90%). The key finding is that the seven highest-frequency indicators (Community Participation, Resilient Planning, Planning Intervention, Governance and e-Governance, Land Use, Resource Management, Urbanisation) appear in the Vital tier under all three threshold settings, confirming that the 80% choice does not alter the core of the framework. The two-track Vital/Useful structure means that even indicators that fall outside the 80% Vital tier are not automatically excluded; they enter the Useful tier and remain available for Stage 4 expert reinstatement, which is how Symbiosis of Industries and PPP Modularity entered the final 32-indicator set despite sitting in the Useful tier.

3.4 Stage 4: expert validation using the staticized group technique

The Pareto-selected indicators were subjected to expert validation through the Staticized Group Technique (SGT), a structured single-round expert elicitation technique that aggregates individual domain judgements excluding the iterative consensus pressure inherent in Delphi approaches (Ulschak, 1983; Graefe and Armstrong, 2011). An expert panel of 18 domain specialists was purposively selected across five professional domains: Urban and Environmental Planning, Transport and Supply Chain Management, Urban Finance, Public Health, and Disaster Risk and Governance (Figure 3). Purposive sampling ensured all panel members possessed expertise in their domain directly, with a minimum of 5 years of professional experience or at least five publications in recognised journals. The structured survey was administered from February to April 2025.

All 18 experts participated voluntarily and provided written informed consent prior to completing the survey instrument. Consent was obtained either via a signed consent form or secure written confirmation. The survey did not collect any personally identifiable information; responses were coded anonymously using participant identifiers (E01–E18). Ethical details are provided alongside the Ethics Statement mentioned below.

Expert feedback resulted in the elimination of two indicators (Optimised Land Use, subsumed by Heterogeneous Resilient Planning; and Coastal Resilience, subsumed by Resilient Infrastructure Coping Capacity), the merging of five further indicators into existing categories (Biodiversity Index into Green Infrastructure; Digital Economy and Polycentric Governance into Governance and e-Governance; Renewable Energy Infrastructure into Energy Supply; and the generic Supply Chain Management indicator into Poor Supply Chain Management Practices), the addition of two new indicators (Symbiosis of Industries and PPP Modularity), and nomenclature refinements applied to several retained indicators for terminological precision. The final validated indicator set comprises 32 RIs, derived from the 37 Vital indicators carried forward from Stage 3, less the two eliminations, less the five mergers, plus the two additions (37–2 - 5 + 2 = 32). The full reconciliation of this arithmetic against the Stage 3 Pareto indicator list is provided in the Reviewer Response Log sheet of Supplementary File 2. The SGT was selected over Delphi and nominal group technique for three methodological reasons specific to this study context. The technique has demonstrated applicability beyond the quality-improvement and healthcare contexts in which it was originally developed, including its use in transport-behaviour expert panels assessing barriers and incentives to policy interventions (Rey-Merchán et al., 2022), supporting its suitability for the present indicator-development application. First, the single-round design eliminates the anchoring effect that arises in iterative methods when early-round consensus shapes later-round responses: in a study validating indicators whose weighting will be determined by the same panel, anchoring bias would compound across the AHP pairwise comparison stage, potentially inflating agreement on indicators that received early endorsement regardless of their intrinsic merit. Second, the statistical aggregation of independent judgements in SGT preserves the epistemic independence of expert perspectives, which is particularly important in a purposive panel where panellists from different professional domains may hold systematically different views about which urban resilience dimensions are most important. Third, single-round administration is more practical for a geographically dispersed panel: the 18 specialists were distributed across multiple cities and institutions, and a multi-round process would have introduced attrition risk and extended the data collection timeline beyond the February to April 2025 administration window. The elimination of Optimised Land Use and Coastal Resilience, the merger of five further candidate indicators into existing categories, and the addition of Symbiosis of Industries and PPP Modularity, by the expert panel illustrate the value of Stage 4 validation: both added indicators reflect emerging governance and supply chain concerns that the literature base, dominated by pre-2022 publications, had not yet fully registered as resilience priorities. Last-Mile Connectivity, by contrast, was retained by the panel as a distinct indicator from Accessibility, on the grounds that last-mile service gaps in peripheral and informal settlements represent a qualitatively distinct resilience constraint from city-wide accessibility metrics, even though the two indicators are conceptually related.

3.5 Stage 5: prioritisation via analytic hierarchy process (AHP)

The Analytic Hierarchy Process (AHP), developed by Saaty (1980), transforms qualitative judgements into quantitative priority weights through pairwise comparison matrices. Multi-criteria decision-making techniques, including AHP, have been widely applied for ranking and prioritisation tasks (Jahan et al., 2016). The AHP model is structured across three levels: (a) objective, overall urban resilience; (b) criterion, the seven dimensions; and (c) sub-criterion, the 32 individual indicators. A fourth level maps criteria against seven theoretical alternatives from urban systems theory (Figure 4). Analysis was conducted using XLSTAT (version 2024.4.0.1424). Expert matrices were aggregated using the geometric mean method. Consistency was verified using the Consistency Ratio (CR), with the threshold of CR at or below 0.10 per (Saaty, 1980). The procedural choices made at each stage of the AHP implementation, problem modelling, pairwise comparison administration, judgement scale selection, weight derivation, consistency checking, and sensitivity analysis, follow the methodological standards set out in the most widely cited review of AHP practice (Ishizaka and Labib, 2011). The geometric mean method was selected over the arithmetic mean for aggregating expert matrices because it is the mathematically appropriate aggregation procedure for ratio-scale data: AHP pairwise comparisons are ratio judgements, and the geometric mean of ratio-scale judgements produces a ratio-scale aggregate, whereas the arithmetic mean does not (Grošelj et al., 2015). This distinction has practical consequences when expert judgements diverge substantially: the arithmetic mean is sensitive to outliers in a way that the geometric mean is not, and in a panel where one panellist assigns an extreme priority to a dimension that others rate moderately, the arithmetic mean will distort the aggregate weight toward that extreme. The AHP hierarchy’s fourth level, mapping criteria against seven theoretical alternatives from urban systems theory (Urban Flow, System Resilience, System Evaluator, Forces, System Characteristics, Action Type, and Community), serves a verification function rather than a weighting function: it tests whether the relative weights assigned to the seven dimensions are consistent with the relative importance of the theoretical constructs those dimensions are designed to operationalise. Urban Flow emerges as the dominant alternative (11.05%), confirming that the highest-weighted dimensions collectively capture the metabolic circulation of resources, people, and information that is central to urban systems theory. This cross-validation of the AHP weight structure against theoretical alternatives provides an additional layer of validity evidence beyond the Consistency Ratio check.

4 Results

4.1 Final 32 resilience indicators

The five-stage process yielded a final set of 32 Resilience Indicators distributed across seven dimensions. Table 2 presents the complete indicator set with AHP-derived weights.

TABLE 2. Final 32 resilience indicators with dimensions and AHP priority weights.

Source: XLSTAT AHP synthesis output, geometric-mean aggregation across the 18-member expert panel (Supplementary File Sheet S5a). Dimension weights verified to sum to 100.00%

4.2 Dimension priority weights

Table 2 presents the complete AHP-derived weights for all seven dimensions and their 32 constituent indicators, discussed further below.

4.2.1 Alternative priority weights

Urban Flow emerges as the most significant theoretical alternative (31.33%), reflecting its capacity to capture interconnected dynamics of people, goods, energy, water, and information constituting the metabolic basis of urban resilience. System Resilience ranks second (20.51%), emphasising redundancy, robustness, and resourcefulness. Community ranks lowest (2.09%), reflecting its role as an outcome of resilience, with sub-indicators embedded within the USP and HWB dimensions. The alternative weight distribution is internally consistent with the dimension priority structure. Urban Flow (31.33%) dominates because the top two dimensions—Urban Design and Spatial Planning and Disaster Preparedness—together account for 53.43% of total framework weight. Urban Design and Spatial Planning governs the arrangement and connectivity of urban space, while Disaster Preparedness governs the resilience of the infrastructure networks—transport, water, energy, and information systems—that keep those spaces functioning under stress. Urban Flow accordingly captures the combined imprint of spatial organisation and infrastructural robustness on the metabolic circulation of people, goods, water, energy, and information through the city. System Resilience (20.51%) maps onto the combination of Disaster Preparedness and Ecosystem and Resources, which together address the structural robustness and redundancy of urban provisioning and hazard management systems. The lower weights of Forces (11.80%) and System Characteristics (9.40%) reflect expert consensus that the driving pressures on urban systems and their structural properties are important contextual factors but secondary to the planning, infrastructure, and governance dimensions that are the direct instruments of resilience intervention. Community (6.15%), despite its theoretical centrality in resilience discourse, registers the lowest alternative weight because the social cohesion and participatory governance dimensions it represents are captured within HWB and USP rather than standing as independent systemic forces in the ARISE framework. This positioning is consistent with the UPE theoretical grounding of ARISE, which treats community-level resilience as a product of equitable spatial and institutional arrangements rather than an independent driver of urban system performance.

4.3 Consistency validation

Consistency Ratios (CR) were calculated for all 18 panel members at both criterion and sub-criterion levels. Sixteen of eighteen experts achieved CR at or below 0.10 across all comparison matrices. Two panel members (E09, Urban Finance; E14, Public Health) recorded sub-criteria CR values of 0.12 and 0.11 respectively; their judgements were excluded from sub-criteria geometric mean aggregation but retained at the criteria level. The hierarchical consistency of criteria and sub-criteria weights confirms the internal validity of the AHP model. The 88.9% compliance rate (16 of 18 panellists meeting the CR threshold at all levels) compares favourably with published AHP studies in urban resilience and sustainability contexts, where compliance rates between 70% and 85% are commonly reported for panels of comparable size (Haque et al., 2022; Pei et al., 2019). The partial exclusion of E09 and E14 at the sub-criterion level follows standard AHP aggregation practice: their criteria-level judgements, which met the threshold, contribute to the dimension priority structure, while their sub-criteria judgements, which marginally exceeded the threshold, are excluded from the indicator-level weighting. The effect of this partial exclusion on the final weight structure was assessed through the leave-one-out sensitivity analysis reported in Section 6, which confirms that the exclusion of E09 and E14 does not materially affect the stability of the six highest-weighted dimensions.

4.4 Empirical data availability analysis: five indian metropolitan cities

To respond directly to the question of whether ARISE’s 32 indicators can be populated from data that currently exists in the public domain, a systematic documentation search was conducted across five Indian metropolitan cities: Chennai, Mumbai, Delhi (NCT), Bengaluru, and Kolkata. This was not an attempt to apply ARISE as a validated assessment tool, an exercise that, as discussed throughout this paper, would require primary data collection, expert re-calibration for each local context, and a measurement and scoring protocol that does not yet exist. Rather, the purpose was narrower and more tractable: to determine, indicator by indicator and city by city, whether a public secondary source containing relevant data could be identified, and to analyse the pattern of where such sources exist and where they do not. This constitutes a data-availability audit rather than a resilience assessment, and the distinction is maintained throughout the analysis and its interpretation.

The search drew on four document categories: city climate action and resilience plans (Chennai 2025, Mumbai 2022, Delhi 2021, Bengaluru 2024, Kolkata 2025); see also Ahmad and Hassan (2025) on nature-based solutions for resilient and sustainable cities in India; multilateral development bank project records (World Bank, ADB, IFC); government audit and statutory reports (CAG 74th Amendment audit, Economic Survey of Delhi); and independent survey-based sources (OMI Foundation, 2022). Full source citations are in Supplementary File Sheet S6.

Of the 160 indicator-city combinations (32 indicators across 5 cities), 85 (53.1%) had at least some identifiable public secondary documentation; 75 (46.9%) had none. At the indicator level, 28 of the 32 indicators (87.5%) had documentation for at least one city, while four indicators, Spatial Network and Regulations, Disaster Preparedness Response and Recovery Management, Stabilised Resource Network, and Symbiosis of Industries, had no identifiable public source for any of the five cities. At the city level, Chennai had the most complete documentation (23 of 32 indicators, 71.9%), followed closely by Kolkata (22 of 32, 68.8%), with Bengaluru (16 of 32, 50.0%), Delhi (15 of 32, 46.9%), and Mumbai (9 of 32, 28.1%) trailing. This city-level spread, however, did not reach conventional statistical significance in a chi-square goodness-of-fit test against the null hypothesis of equal documentation across cities (χ² = 7.65, df = 4, p = 0.105), indicating that the apparent gap is consistent with chance variation at this sample size rather than a robust, generalisable difference between cities.

Documentation rates varied by dimension (Economic Environment 87% to Infrastructure and Logistics 37%), but chi-square tests found no statistically significant dimension-level pattern (χ² = 5.03, df = 6, p = 0.540; Cramér’s V = 0.18) — the spread is consistent with chance variation at this sample size. The pattern is better explained by which document types exist (climate action plans cluster around emissions and disaster indicators; CAG audits cluster around governance) than by any dimension-specific bias.

We explicitly did not compute a composite ARISE score or city ranking from this dataset. With only 53.1% of indicator-city cells documented, and populated cells spanning incompatible formats, any composite score would reward documentation completeness rather than actual resilience. This section maps where India’s current urban data infrastructure can and cannot support systematic resilience assessment.

The 46.9% documentation gap is not randomly distributed, and its pattern is informative rather than merely a limitation to note. Consistent with the non-significant dimension-level chi-square test reported above, the gap tracks which public document types exist for a given city rather than any single dimension being institutionally neglected. Infrastructure and Logistics indicators (37% documentation, the lowest of any dimension) depend heavily on asset-specific engineering records, such as stormwater drainage capacity or transport network redundancy, that Indian municipal and parastatal agencies typically hold internally rather than publish, even where the underlying infrastructure exists and functions adequately. Economic Environment indicators (87%, the highest), by contrast, draw on macroeconomic and census-linked data that state and national statistical agencies already publish on a fixed release cycle regardless of any individual city’s planning capacity. This suggests the primary constraint on future ARISE deployment is less about any single dimension being neglected, and more about the uneven maturity of open-data mandates across different categories of urban record-keeping: macroeconomic and climate data are increasingly standardised for public disclosure, while asset-level infrastructure and institutional-process records remain fragmented across agencies with no uniform publication requirement.

5 Discussion

5.1 ARISE as a UPE-Grounded theoretical contribution

The integration of Urban Political Economy perspectives into the ARISE indicator development process constitutes the framework’s primary theoretical contribution. Unlike existing frameworks that treat urban resilience as a politically neutral technical property, ARISE treats resilience as a politically mediated outcome shaped by the distribution of power, capital, and institutional authority across urban space. This is achieved through three functional mechanisms. First, the method for indicator selection criteria clearly prioritises indicators that captures irregularities in governance mechanisms (Power Devolved to LPA), consistent with participatory governance mechanisms documented in comparable South Asian contexts (Clay, 2007), politicised resource allocation (Politicised Resource Flow), and distributional inequality (Geographical Inequality, Inner-City Poverty). Second, urban metabolism, the key UPE notion of energy, water, material, and nutrient fluxes through urban systems, has a direct impact on the Ecosystem and Resources dimension. Third, the Circular Economy Potential and Industrial Symbiosis highlight UPE’s commitment to evolving beyond linear economic frameworks. Meerow and Newell’s (2019) demand for resilience frameworks that address the distributional and political aspects of urban vulnerability is explicitly addressed by these contributions. The UPE foundation of ARISE has a second theoretical implication that goes beyond the particular indicators: it shifts the goal of resilience assessment. ARISE questions if the pre-disturbance equilibrium was equitable in the first place, whereas dominant frameworks quantify the extent to which a city system may return to that equilibrium. An urban region that quickly recovers its pre-existing spatial inequities, infrastructure shortfalls, and governance exclusions following a catastrophe is robust in engineering terms, but not in social justice terms. The normative stance that resilience must be assessed against distributional criteria rather than just systemic continuity is encoded by ARISE, which weights Geographical Inequality and Politicised Resource Flow as high-priority indicators. This adds to the expanding corpus of study, which includes (Meerow and Newell, 2019) and Sharifi (2023), that questions whose resilience is being generated and at what cost.

5.1.1 Reconciling the equity orientation of ARISE with its empirical weight structure

A tension in the ARISE weight structure deserves explicit acknowledgment rather than passing mention. The framework is theoretically grounded in Urban Political Economy and presents itself as an equity-oriented instrument, yet the empirical weights produced by the expert panel place Ecosystem and Resources — the dimension operationalising UPE's urban-metabolism concept — at only 5.96% of total framework weight, the lowest of the seven dimensions. Governance fares only slightly better at 8.62% (fifth of seven), and its constituent indicators include the single lowest-weighted indicator in the entire 32-indicator set: PPP Modularity (0.69%), followed closely by Community Participation (0.75%, Health and Wellbeing) and Power Devolved to LPA (1.02%, Governance). Urban Design and Spatial Planning, by contrast, receives 36.04%, more than six times the weight of Ecosystem and Resources and more than four times the weight of Governance. This pattern — the empirical marginalisation of the two dimensions (Ecosystem and Resources, Governance) that most directly operationalise UPE's distributional and political concerns — requires critical examination rather than convenient reconciliation with the framework's stated theoretical orientation.

First, panel composition: none of the panel's five professional domains is a dedicated ecological or natural-resource specialism — ecosystem-relevant expertise is subsumed within the broader ‘Urban and Environmental Planning’ category (38.9% of the panel), alongside spatial-planning expertise, while governance and disaster-risk specialists together comprised only 5.6%. Leave-one-out sensitivity analysis (Supplementary File Sheet S5a) confirms Ecosystem and Resources is by far the most panel-sensitive dimension (range 4.36%–7.74% across 18 iterations, CV 22.6%), more than twice as volatile as Governance (range 7.55%–10.15%, CV 12.5%) and Health and Wellbeing (range 5.54%–7.91%, CV 14.8%) — the two next most sensitive dimensions. The subgroup analysis (Section 5.1.1.1; Supplementary File Sheet S5b) confirms USP dominates in all three professional subgroups, and that Ecosystem and Resources ranks last in all three, including the Finance/Health/Governance cluster (5.46%–6.07% across subgroups) — so its low weight is not an artefact of any single subgroup's composition, but is consistent with no subgroup containing a dedicated ecological-resource specialist. Governance likewise ranks second-lowest in all three subgroups (8.16%–8.89%), including the Finance/Health/Governance cluster that contains the panel's sole disaster-risk-and-governance specialist — indicating its low weight is not simply a case of governance being unrepresented either.

Second, more speculatively: domain experts may treat both ecological-resource management and governance capacity as causally downstream of spatial planning—contingent on whether the physical and infrastructural conditions for effective resource stewardship and institutional coordination are first in place. This is a hypothesis consistent with the sequencing logic in Section 5.3.

Third, and most fundamentally: the low weights on Ecosystem and Resources and Governance alike may reflect a structural limitation of expert-based AHP for equity- and sustainability-oriented frameworks. Domain experts may systematically undervalue ecological stewardship and participatory, devolutionary governance mechanisms regardless of panel composition (Krick, 2022). ARISE should therefore be read as an expert-prioritisation framework whose weight structure reflects professional judgements, not participatory community priorities or ecological carrying-capacity constraints. The priority weights in Table 2 do not incorporate the lived experiences of informal settlement residents, vulnerable populations, or community organisations, nor an independent ecological assessment—a first-order limitation identified as a next research phase in Section 6.

5.1.1.1 Subgroup weight comparison by professional domain

To address the concern about field-specific bias directly, rather than only acknowledging it qualitatively, a subgroup analysis was conducted comparing the dimension-level priority weights derived from three subgroups of the 18-member panel: (a) planners and environmental specialists (n = 7, 38.9%), who might be expected to assign higher weight to Urban Design and Spatial Planning; (b) transport and supply chain specialists (n = 5, 27.8%), who might be expected to assign higher weight to Infrastructure and Logistics; and (c) the remaining six panellists spanning Urban Finance, Public Health, and Disaster Risk and Governance (n = 6, 33.3%). The geometric mean was computed separately for each subgroup across the seven dimension-level pairwise comparisons, producing three independent weight vectors, which are compared against the aggregate weight vector in Supplementary File Sheet S5b.

Supplementary File Sheet S5b reports criteria-level weights, computed independently from the 7 × 7 dimension pairwise matrix for the full panel and for each professional subgroup. These agree closely with the global dimension weights in Table 2 (e.g. IL: 10.57% criteria-level vs. 10.58% global; USP: 36.20% criteria-level vs. 36.04% global), confirming the two levels of the AHP hierarchy are internally consistent. The subgroup comparison in Table 5b's underlying data shows Ecosystem and Resources ranking last in all three subgroups (5.46%–6.07%) and Governance second-lowest in all three (8.16%–8.89%), so neither dimension's low weight can be attributed to the particular mix of specialisms on the full 18-member panel.

5.1.2 Answering the central question: What does UPE-Informed thinking add?

Section 5.1.2 answers the paper’s central question—what UPE-informed thinking adds—with three specific, checkable claims rather than a general assertion of theoretical relevance.

A counterfactual process without UPE would plausibly produce sector-organised dimensions (water, energy, transport) rather than political-economic ones (resource flow, spatial inequality, governance devolution). Indicators such as Politicised Resource Flow, Geographical Inequality, and Power Devolved to LPA enter ARISE because UPE identifies them as constitutive of resilience outcomes, not because they were frequent in the literature pool.

Second, UPE motivated a specific procedural safeguard at Stage 3: the two-track Vital/Useful retention structure exists precisely because a UPE-grounded framework cannot rely on citation frequency alone without contradicting its own theoretical premise. If political and distributional indicators are under-represented in mainstream resilience literature (as UPE argues), a purely frequency-driven Pareto cut would systematically eliminate the indicators the theory identifies as most important. Symbiosis of Industries and PPP Modularity, both added at Stage 4 rather than surviving the Pareto cut, are concrete evidence that this correction functioned.

Third, UPE supplies the interpretive standard against which ARISE’s own AHP weighting outcomes are judged. The low weights on Ecosystem and Resources (5.96%) and Governance (8.62%) are only identifiable as a problem requiring explanation because UPE supplies an external criterion—political, distributional, and ecological centrality—against which to judge the empirical result. A theory-free framework would have no principled basis for treating that weight as anything other than a neutral empirical output.

5.2 Methodological contribution

This research extends resilience indicator methodology through three specific advances: (1) the sequential combination of Pareto analysis and SGT validation provides a more structured reduction approach than single-round expert surveys or purely subjective selection (Sharifi, 2016; Bruckler et al., 2024); (2) UPE serves as an explicit classification criterion rather than background narrative, offering theoretically grounded justification for the categorisation structure; and (3) the documented interrogation of AHP weighting outcomes against theoretical commitments is a procedural contribution transferable to other theory-grounded indicator frameworks.

5.3 Practical implications

Table 3 translates the highest-priority dimensions into expert-recommended planning implications, ordered by dimension weight as a provisional prioritisation sequence. These implications reflect expert professional judgements, not validated causal evidence; they must be contextualised to each city’s regulatory environment and governance capacity before informing resource allocation decisions.

TABLE 3. Policy and planning implications of priority-weighted resilience dimensions.

Source: Author.

To make these priority-weighted recommendations concrete for practitioner audiences: consider a data-constrained metropolitan city facing recurring monsoon flooding in informal settlements, a scenario documented across several of the cities audited in Section 4.5. Under ARISE, this scenario would directly activate at least three indicators spanning three dimensions: Resilient Infrastructure Coping Capacity (DP) would flag inadequate stormwater drainage and emergency-response readiness; Green Infrastructure and Community Participation (HWB) would flag the absence of blue-green infrastructure and resident consultation in the affected settlement; and Resource Allocation and Accessibility (IL) would flag whether post-disaster infrastructure investment reaches the affected area proportionally to its exposure. Read together across dimensions rather than in isolation, these three indicators would identify not only that flood risk exists, but which of governance, infrastructure, or participatory planning gaps is driving it in a given city, illustrating how ARISE’s cross-dimensional structure is intended to support diagnosis rather than only single-issue risk-flagging.

6 Limitations and future research

Several limitations qualify the scope and generalisability of this study. First, the 18 panellists' professional settings are reflected in the priority weights given to each dimension. Leave-one-out resampling (18 iterations; Supplementary File Sheet S5a) shows the four highest-priority dimensions—USP (CV 5.7%), DP (CV 4.9%), EE (CV 8.4%), and IL (CV 8.3%)—are stable, with coefficients of variation below 10%. Governance (8.62%) is moderately sensitive (CV 12.5%, range 7.55%–10.15%), and both Health and Wellbeing (6.56%; CV 14.8%, range 5.54%–7.91%) and, most markedly, Ecosystem and Resources (5.96%; CV 22.6%, range 4.36%–7.74%) are considerably more panel-sensitive. This ordering is consistent with panel composition: none of the five professional domains on the panel is a dedicated ecological or natural-resource specialism, and governance/disaster-risk specialists comprised only 5.6% of the panel. Additional ecological-resource and governance domain experts should be added to the expert panel when using ARISE in ecologically fragile or governance-constrained environments respectively. Concretely, we propose two steps for the next research phase: a targeted second-round Delphi consultation restricted to governance and public-health specialists, recruited specifically to reach parity with the two largest existing subgroups (planning and environmental specialists, n = 7; transport and supply chain specialists, n = 5), so that a re-weighted, representation-balanced AHP result can be directly compared against the present panel’s output; and reporting both weight sets side by side in any future application, so that the sensitivity of the Governance weight to panel composition is made visible to readers and city practitioners rather than left implicit. Second, this study offers a cross-sectional evaluation; future research should prioritise long-term tracking of indicator weights. Third, data availability differs depending on the type of city, even though the prototype is intended for data-constrained urban applicability. Fourth, the framework has not yet been tested through an urban case study; the next stage of this research program will empirically validate the framework through application to a real city scenario, with a focus on South or Southeast Asian contexts. Of the five cities audited in Section 4.5, Chennai is proposed as the specific candidate for this first pilot, since it returned the highest documentation completeness of the five (23 of 32 indicators, 71.9%), reducing the risk that an initial proof-of-concept scoring exercise would be confounded by data unavailability rather than reflecting genuine resilience conditions. Such an application would provide the empirical foundation that this methodology study lays out, enabling the framework to be assessed not just against expert opinion but also against observable resilience results across a real city’s seven-dimensional drivers. Through case study deployment, the gap between a validated indicator framework and an applied diagnostic tool is filled, and ARISE is built to be deployable in situations without extensive data systems. Future studies should investigate participatory extensions of the framework that take into account the priorities and expertise of disadvantaged urban populations, whose lived experiences of urban vulnerability are not always adequately represented by the opinion of domain experts alone. It is the responsibility of a framework whose theoretical underpinnings are found in urban political economy to compare its commitment to equality with the viewpoints of people whose equity is at risk.

Fifth, measurement protocol specification, missing-data treatment rules, and cross-city differentiation protocols remain unresolved—necessary preconditions for empirical application that are identified as priority next steps. Two methodological pathways are proposed as candidates for this phase: Coupling Coordination Degree analysis (Sutar et al., 2022), to test whether improvements in one dimension, such as Urban Design and Spatial Planning, reliably coincide with improvements in a related dimension, such as Infrastructure and Logistics, rather than assuming dimensions operate independently; and an AHP-based composite index approach (Sutar et al., 2023), which offers a template for aggregating ARISE’s 32 indicators into a comparative resilience score, including specific normalisation techniques (Min-Max scaling or Z-score standardisation) and missing-data treatment rules suited to the heterogeneous, incomplete data documented in Section 4.5.

7 Conclusion

ARISE's documented divergence between the theoretical centrality of governance and ecological-resource stewardship within Urban Political Economy, and their low empirical weights—Ecosystem and Resources (5.96%) is the lowest-weighted dimension overall, and Governance (8.62%) contains the single lowest-weighted individual indicator (PPP Modularity, 0.69%)—is the clearest evidence that theory and method operated independently before being combined. This study is an indicator-development and expert-prioritisation framework, not a validated urban assessment tool; ARISE has not yet been applied to a real city, and measurement protocol specification and case-study validation are the next research phases.

Future research should pursue empirical validation through application to specific urban case studies in South and Southeast Asia, where the source literature is most densely concentrated and where South and Southeast Asian urbanisation pressures are most acute. Participatory extensions of the ARISE methodology, incorporating the knowledge and priorities of marginalised urban communities alongside domain expert judgement, would address the most significant remaining gap in the framework’s claim to equity-orientation: the current expert panel, however purposively selected, does not include the lived-experience perspectives of those whose urban vulnerability the framework is designed to assess. Temporal modelling to track how resilience indicator weights shift as urban conditions, governance arrangements, and ecological pressures evolve over development phases would convert ARISE from a cross-sectional diagnostic into a longitudinal monitoring instrument. Finally, the Governance dimension’s sensitivity to panel composition warrants a dedicated study examining how the relative weight of institutional governance in urban resilience assessment varies across governance systems of different types, including federal and unitary states, strong and weak local government traditions, and varying degrees of public-private partnership culture. These four research directions collectively constitute the programme through which ARISE can be developed from a validated methodological framework into a field-tested, continuously refined instrument for equitable urban resilience governance.

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/fenvs.2026.1889318/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Esther Oreofeoluwa Esho, The Australian New Zealand Society for Ecological Economics (ANZSEE), Australia

Reviewed by: Hamid Awad, University of Anbar, Iraq

Goutam Sutar, Thiagarajar School of Management, India