Introduction

Flood risk is increasing globally due to climate change. Intense rainfall, variable hydrological regimes, and rising sea levels are escalating compound coastal and pluvial flood events. Rapid urban expansion into flood-prone areas has increased the exposure of assets and vulnerable populations. Currently, approximately 1.6 billion people reside within the 1% annual-chance floodplain, a figure projected to rise to nearly 1.9 billion by 2100. Effective flood risk governance requires a framework that considers flood physics, makes uncertainty explicit, and distinguishes between avoidable error and systematic bias. This is particularly vital for data-scarce regions where limited observations and institutional capacity hinder infrastructure planning and early warning effectiveness.

Developing The Framework

The proposed multilevel flood risk governance framework integrates Hazard modelling (H), social Vulnerability assessments (V), and risk Communication (C). Guided by a constructivist epistemology, the framework views flood risk as a co-produced outcome of physical processes, social structures, and communicative infrastructures. This approach responds to long-standing criticisms regarding the fragmentation of flood modelling and communication efforts, which often operate in disciplinary silos. The framework is designed to be adaptable for regions with limited data and high climate vulnerability, prioritising action under limited data and capacity.

Foundational Synthesis And Identification Of Gaps

A review of peer-reviewed articles and global indices reveals operational limitations in existing flood management. Hazard modelling often treats vulnerability as an external input rather than a core component. Vulnerability assessments frequently rely on national-scale composite indices, overlooking micro-spatial dynamics such as informal settlements. Communication tools are often developed separately from the models, resulting in unclear messaging and inaccessibility for marginalised groups. The framework seeks to create interoperability between these domains to improve decision-making, particularly where data and institutional coordination are limited.

Deconstruction And Boundary Analysis

The framework disaggregates each domain into operational components. Hazard metrics include flood depth, duration, flow velocity, and return periods. Social vulnerability considers indicators such as age, income, and housing type. Communication involves target groups, message formats, and delivery mechanisms. Three key boundary tensions are identified: H-V (misalignment in scale), H-C (unfriendly statistical reporting), and V-C (barriers for marginalised groups). The framework addresses these by prioritising the finest resolution data on vulnerability and using layered, user-facing communication formats.

Indicator Selection And Rationale

Indicators are selected based on relevance to flood risk, data availability, and applicability under resource constraints. For H, maximum flood inundation depth and annual exceedance probability (AEP) are primary. The V component includes socioeconomic status, demographic risk, and housing conditions. For C, indicators capture access and interpretation, such as mobile reach and trust in local institutions. Proxy measures, such as satellite-derived settlement footprints or nightlight intensity, can be used in data-scarce or informal contexts to ensure the framework remains globally comparable.

Design Principles

Three principles guide the framework’s integration. First, joint interpretation ensures hazard metrics are interpreted alongside vulnerability indicators. For instance, 1m of flood depth is more dangerous in informal dwellings than reinforced housing. Second, clear representation of uncertainty uses multi-format communication, such as shaded probability zones and verbal descriptors. Third, functionality under data scarcity ensures the framework operates with minimal inputs, such as coarse digital elevation models (DEMs) at 10–30m resolution, where higher-resolution data are not available.

Presenting The Framework

The framework is presented as a Venn diagram with a central decision-ready space. H denotes physical flood characteristics, V explains unequal consequences across groups, and C turns knowledge into signals, visuals, and procedures. Synthesis happens at the interfaces: H-V for consequence-focused ranking, H-C for impact-based warnings, and V-C for tailored messaging. The central decision space explicitly represents uncertainty and matches options to responsible actors, responsibilities, and triggers to ensure feasibility constraints are transparent.

Interfaces Pairwise Intersections

At the H-V interface, the framework ranks places by expected consequences rather than hazard alone. This identifies ‘hot spots’ where a flood scenario would cause disproportionate impacts. The H-C interface translates model outputs into actionable products like impact-based warnings and arrival time windows. The V-C interface aligns content and channels with the capacities of target groups. Finally, the triple intersection H-V-C creates the decision-ready space where forecast, impacts, and actions are integrated for non-technical audiences, specifying ‘who does what, when, and why’.

Operationalising The Framework Theoretical Scenario

The framework is applied to a theoretical 1% AEP rainfall-driven flash/pluvial flood scenario. Hazard screening begins with an open-access DEM at the finest resolution available, typically 10–30m. Social vulnerability is assembled using minimum viable indicators like poverty proxies and age structure. These are combined to produce ‘consequence surfaces’ for spatial prioritisation. Communication strategies are tailored to preferred channels, such as radio or printed maps for households without mobile data. This allows municipalities with minimal capacity to initiate effective action planning through threshold-action matrices.

Policy Insights

Policymakers are encouraged to treat communication as a core component of risk governance rather than a final dissemination step. The framework enables coordination across governance levels, allowing local governments to use proxy indicators while national agencies use complex modelling. It supports iterative learning through regular updates of risk surfaces and decision logs. By applying IRGC risk governance principles, it makes uncertainty visible and reduces the risk of false precision in contested decisions. This guides decisions under conditions of deep uncertainty by making assumptions and alternative pathways explicit.

Conclusion

The multilevel H-V-C framework provides a structured system for translating technical analyses into decision-relevant outputs. It is particularly valuable in data-scarce environments for producing consequence-oriented risk surfaces and transparent decision protocols. By embedding communication within the core of flood risk assessment, the framework enables better preparedness and response under non-stationary climatic and social conditions. Future work should focus on testing the real-world applicability of this framework in municipalities with different levels of data availability and assessing the robustness of interfaces under rapid-onset scenarios.