Multi-hazard risk reduction is increasingly recognized as essential for effective disaster risk management, yet its operationalization remains constrained by decision-making frameworks not designed for cross-hazard comparison. This paper presents a unified microscale framework for harmonized flood and earthquake risk assessment and multi-hazard intervention appraisal in residential buildings. Here, microscale refers to building-specific assessment in which hazard intensity, exposure attributes, vulnerability functions, and intervention appraisal are represented at the level of individual residential buildings rather than aggregated spatial units. The framework establishes a common structure for hazard representation, exposure characterization, vulnerability modelling, and decision variables across flood and seismic domains while retaining hazard-specific physical models, enabling consistent evaluation of single-hazard as well as multi-hazard intervention strategies.The methodology combines building-level flood and seismic risk assessment with multi-dimensional performance metrics, including expected losses, life-safety outcomes, habitability, and environmental impacts. Flood and earthquake risks are evaluated using hazard-specific models, while their annualized consequences and intervention outcomes are brought together at the multi-hazard appraisal stage. Intervention appraisal is supported through feasibility screening based on payback-period constraints and Pareto-efficient analysis to identify non-dominated retrofit options prior to a stakeholder-driven decision. The framework is demonstrated through an application to a portfolio of residential buildings in Pesaro (Italy) exposed to riverine flooding and seismic hazard under multiple hazard and vulnerability scenarios.Results show that changes in seismic demand, building vulnerability, and payback-period thresholds reshape the intervention decision space, shifting the portfolio from a feasibility-limited regime dominated by no action or flood-only measures toward broader adoption of integrated strategies. In the investigated case study, under lower seismic demand, adoption of interventions is limited, and integration emerges primarily as an extension of economically viable single-hazard measures. As seismic demand and vulnerability increase, integrated strategies become structurally competitive, expand the actionable portfolio, and significantly shorten payback periods through cost sharing and synergy effects, although benefits remain typology-dependent. More generally, the proposed framework provides an operational pathway for translating multi-hazard principles into building-scale decision support and offers practical insights for planners, asset managers, and policymakers seeking coherent and cost-effective risk reduction strategies.
Studying human consequences caused by landslides and floods is important to understand their impact and to investigate public risk perception We analysed a catalogue of 1,188 landslide and 752 flood fatalities that occurred in Italy during the 60-year period 1965–2024, for which information on sex, age, and circumstances of death was available. Observed fatalities were compared with expected distributions modelled from census data by sex and age for two non-overlapping 30-year periods. Observed fatalities were compared with expected distributions modelled from census data by sex and age for two non-overlapping 30-year periods. With the exception of children, male fatalities were consistently over-represented relative to the census-based demographic expectations, whereas female fatalities were generally under-represented. To investigate possible relationships with risk perception, we analysed responses from four nationwide surveys conducted in 2012, 2013, 2015, and 2023. Women reported slightly higher levels of self-perceived exposure to natural and technological risks than men, whereas no substantial gender differences emerged in the perceived threat posed by landslides and floods to personal safety. The national-scale analysis was complemented by a regional comparison between perceived threat and historical mortality. Regional Threat Indices were calculated separately for floods and landslides across the 20 Italian regions. A moderate and statistically significant positive association emerged between perceived flood threat and flood mortality, whereas no association was found for landslides. Overall, the results highlight a persistent mismatch between gender patterns in mortality and perceived personal threat and support targeted, gender-sensitive risk communication and disaster risk reduction strategies.
Greater wildfire frequency and severity have increased the need for disaster preparedness and related education, yet household disaster preparedness and response, which remain low, rely on accessible information. Such information is limited in rural communities of the Great Plains (GP) of the U.S that lack resources. We examine the influence of alternative information sources on households’ perceived level of wildfire and disaster preparedness in a wildfire context and perceived confidence in being prepared for a wildfire and disaster, comparing households in the GP and mountainous western (MW) regions of the U.S. We analyze three years of annual household disaster preparedness survey data collected by U.S. Federal Emergency Management Agency, focusing specifically on households surveyed about wildfires. We find that the perceived level of wildfire and disaster preparedness, or perceived confidence level to prepare for a wildfire or disaster, does not differ significantly between the two regions, but prior wildfire experiences and the influence of disaster preparedness information sources shape the perceived level of wildfire and disaster preparedness and confidence in being prepared differently for households in the two regions. Public announcements served as the most influential information source in MW, while network sources served as the most influential in the GP, even for the adoption of defensible space, a wildfire mitigation practice. Practical implications from the research suggest targeting more than one area-specific communication channel in a region for disseminating wildfire disaster preparedness information, as well as providing needed resources that ensure effective and efficient distribution of disaster preparedness information.
Background Increasingly frequent and complex emergencies heighten the need for accessible risk communication that improves public safety and upholds the rights and dignity of all community members. This study examined how accessible risk communication is conceptualized and implemented across emergency management organizations, and developed tools to support practice. Methods We conducted eight adapted Structured Interview Matrix consultations (virtual/in-person) with 72 emergency management and public health professionals across all 13 provinces and territories. Results Using inductive qualitative analysis, we identified existing accessibility initiatives alongside persistent challenges related to resource allocation, time and staffing pressures, siloed workflows, inconsistent standards, and fragmented knowledge infrastructure. Across organizations of varying capacity, participants raised questions about how accessibility could be more efficiently and sustainably embedded into routine practice. Two explanatory models emerged. The Costs of Exclusion Model illustrates how reactive approaches to accessibility may create repeated work and downstream resource demands. The Self-Reinforcing Responsive Model (SRRM) represents how investments in Engagement, Awareness, Preparedness, and Culture build organizational capacity and make accessibility progressively easier and less resource-intensive to sustain. Rather than requiring new tools or major spending, this can involve building on existing resources through relationships, shared learning, technology readiness, and passive preparedness. Conclusions Accessible communication can be reframed from an added burden to a resource-conserving disaster risk reduction strategy. To support implementation, the SRRM is accompanied by accessibility indicators and a phased roadmap that organizations can apply according to their existing capacity. Embedding accessibility before disasters shifts emergency management from reactive correction to proactive preparedness.
In the context of growing risk and uncertainty, disaster risk reduction (DRR) policies are increasingly expected to be evidence-informed. This imperative has contributed to the expansion of research and knowledge infrastructures and to the global circulation of disaster discourse. Yet critical questions remain: what counts as evidence in DRR policymaking, who carries and mobilises it, and with what objectives? These questions are particularly salient in multi-level disaster governance arrangements, where knowledge production and policy authority are dispersed across domestic and transnational actors. This paper critically examines the structures, mechanisms, and pathways that shape the uptake of evidence in DRR policymaking. Drawing on the concepts of the knowledge–policy interface and boundary work, we analyse the interactions between knowledge producers, policymakers, and intermediaries. Empirically, the study employs a multi-method case study approach, combining a rapid review, actor mapping, and semi-structured interviews to identify the actors and formal and informal evidence pathways for policymaking for Nepal. We found that in Nepal, evidence use is not primarily driven by endogenous policy demand; rather, it relied significantly on non-state and external actors who push knowledge into policy arenas. This boundary work operates across three interrelated levels: at the micro level, through relational practices that privilege particular actors and forms of expertise; at the meso level, through semi-institutionalised, often donor- and NGO-driven platforms that facilitate knowledge brokering; and at the macro level, within transnational DRR arenas where donor agencies and epistemic communities define and legitimise knowledge in line with dominant risk and resilience paradigms. Across these scales, evidence circulates iteratively, being filtered, reframed, and selectively endorsed before shaping national DRR policy and programme design.
Rapid urbanisation has intensified disaster risk in informal settlements. In South Africa, over two million households live in informal settlements, many in floodplains or high-risk areas. Often the only accessible land for low-income residents, these sites expose households to flood and fire hazards. Flood risk is intensified by climate variability, rainfall, low-lying terrain and poor drainage, while fire risk is shaped by dense layouts and combustible materials. These hazards interact with poor housing, limited services and low incomes, yet remain insufficiently captured by risk assessments.This study examines how flooding and fire are experienced and managed in Khan Road informal settlement, KwaZulu-Natal Province, South Africa. It draws on 160 interviews by trained community researchers using a survey with closed and open-ended questions. The survey captured hazard exposure, housing and service conditions, impacts and responses.The research contributes household-scale evidence showing how multi-hazard risk is experienced within dwellings and shaped by infrastructure failure, settlement morphology and informal response networks. Findings reveal multi-hazard exposure: over half reported flooding, nearly three-quarters fire, and two-fifths both hazards. Respondents linked flooding to terrain and drainage constraints, and fire risk to dense layouts and combustible materials. Unreliable water supply and shared sanitation intensified vulnerability and constrained recovery. Residents relied on neighbours and family for warning, response and support.The findings indicate that disaster risk in informal settlements is co-produced through morphology, infrastructure deficits and social support systems. They show that community-generated data can strengthen disaster risk reduction and upgrading interventions in South Africa and other Global South cities.
Flooding is among the most widespread and costly natural hazards worldwide, with impacts that vary substantially depending on how quickly it develops. This onset speed is recognised as a critical factor shaping the impacts of natural hazards and the challenges faced by affected communities, yet how these differences play out within a single flood event remains underexplored. To address this gap, we examine the 2022–2023 floods across Eastern and Southern Australia, among the most severe recent natural hazards, causing extensive damage and disruption to affected communities. We conducted 52 semi-structured interviews with residents from four flood-affected communities in Victoria and South Australia (May–June 2024); the Victorian communities experienced rapid-onset flooding, while the South Australian communities experienced slow-onset flooding. Findings reveal that while some experiences were similar across conditions, differences in onset speed were associated with variation in communication channel usage, risk perceptions, protective action behaviours, flood impacts, and recovery challenges, though these patterns should be interpreted alongside demographic and contextual differences between the two samples that may also have contributed to the variation observed. Slow-onset flooding, in particular, introduced unique complexities across the emergency response and recovery phase. These insights highlight the need for policy adjustments, in Australia and internationally, that better accommodate the prolonged nature and associated challenges of slow-onset events.
Flood risk is projected to increase under climate change due to more frequent extreme events and expanding human development in hazard-prone areas. While large-scale assessments provide valuable overviews of emerging risk, they often lack the spatial detail and social-environmental integration needed to inform local adaptation planning. This study applies a high-resolution hazard-exposure-vulnerability framework to assess compound coastal–fluvial flood risk and evaluates how methodological choices affect estimates of exposure, economic damage and spatial risk. The framework is applied to the low-lying Broadland catchment in eastern England using UKCP18 convection-permitting climate projections under RCP8.5 to simulate compound flood scenarios for 1990, 2030, and 2070. Exposure is assessed using population and property data, while vulnerability is quantified across 31 social, economic, and environmental indicators statistically weighted via principal components analysis. The results show that methodological choices can substantially influence outcomes. In particular, building depth sampling methods and depth–damage assumptions can alter loss estimates by over 70%, while different population datasets affect estimated human exposure. By using the OpenPopGrid population dataset and the footprint-maximum depth, the estimated economic damages increase from £439 million in 1990 to £718 million in 2070, with risk concentrated in urban coastal locations where exposure and vulnerability converge. By integrating high-resolution climate projections with spatially explicit vulnerability analysis, this study demonstrates how local-scale assessments can provide spatially detailed evidence to inform adaptation planning. The framework is potentially transferable to other coastal and estuarine regions, although its implementation depends on the availability and local suitability of hazard, exposure and vulnerability data.
Increasingly frequent and intense extreme weather events have intensified time-pressured governmental disaster reporting. In Taiwan, each typhoon or heavy rainfall event requires integrating heterogeneous meteorological, hydrological, and operational data from multiple platforms into a standardized government report. This process typically requires 0.5–3 workdays per event and relies heavily on the expertise of senior specialists. To address these challenges, this study proposes an automated disaster response report generation system that combines web scraping, multi-source data extraction, computer vision, and Large Language Models (LLMs) using Retrieval-Augmented Generation (RAG) and Chain-of-Thought (CoT) prompting to produce draft chapters aligned with current reporting standards. The performance of the proposed system was evaluated using the System Usability Scale (SUS), expert assessments of content quality and production efficiency, the Trustworthy Language Model (TLM) framework, and a Ragas-based comparative evaluation across different generation configurations. The system achieved a mean SUS score of 85.5 (SD = 15.2), reduced report preparation time by 73% for typhoon events and 64% for heavy rainfall events, and attained TLM Trustworthiness scores of 0.89 ± 0.04 and 0.95 ± 0.00 for the two evaluated report sections. The Ragas-based comparison further showed that the proposed RAG-CoT configuration achieved the highest Professional Report score of 0.86 ± 0.04 among the evaluated configurations. These findings indicate the feasibility of using an automated, traceable workflow to reduce manual workload, support production efficiency, lower writing barriers through data integration, and produce evidence-grounded draft content for governmental disaster response reporting under the evaluated conditions.
With the increasing frequency of extreme climate events, the flood safety of underground suburban rail transit stations has garnered significant attention. This study aims to investigate the influence mechanisms of multi-dimensional environmental interventions (human guidance, emergency broadcasting, and environmental stress) and individual characteristics on evacuation wayfinding behavior, thereby quantitatively evaluating the evacuation reliability of underground stations. Using a typical underground station of the recently opened Shanghai Airport Link Line as a prototype, an immersive virtual reality (VR) environment was developed to simulate flood disaster scenarios, involving 987 participants who performed evacuation experiments alongside pre- and post-experiment surveys. This research employs binary logistic regression combined with an ensemble learning algorithm (LightGBM) to perform attribution analysis on complex non-linear behavioral decision-making. Key risk factors affecting decisions were identified, and the SHAP (SHapley Additive exPlanations) framework was introduced to conduct global and local sensitivity analyses. The results indicate that human guidance is the core factor in enhancing systemic evacuation reliability. High levels of staff deployment significantly improve participants' evacuation efficiency; furthermore, a significant synergistic effect exists between sufficient manual guidance and detailed emergency broadcasting, which collectively prompts participants to select the optimal safe route. Notably, the sensitivity of path-finding accuracy to environmental stress decreases significantly under strong human intervention. This study fills the gap in flood evacuation research for suburban railways and demonstrates the superiority of large-sample data in the reliability assessment of emergency evacuations. The findings provide important theoretical guidance for developing high-resilience flood emergency plans and optimizing staff deployment under resource-constrained conditions.
On 8 September 2023, a Mw 6.8 earthquake struck Morocco’s High Atlas region, causing nearly 3,000 fatalities and widespread damage across rural, mountainous settlements dominated by traditional and non-engineered construction, and affecting cultural heritage assets. The event triggered landslides that further increased the isolation of affected communities. This study presents findings from the Earthquake Engineering Field Investigation Team (EEFIT) mission. Different radar and optical remote sensing techniques were used to identify areas most affected by earthquake-induced landslides and building damage, guiding the field investigations. A total of 2,671 earthquake-induced landslides were identified; although relatively limited for an event of this magnitude, they significantly impacted road accessibility and emergency response. In parallel, 455 buildings were assessed on site and classified using the EMS–98 scale. Seismic performance was primarily influenced by construction practices and detailing rather than structural typology. Traditional unreinforced masonry buildings exhibited the highest vulnerability, largely due to substandard materials and construction practices, and limited upkeep, which are common in remote villages. Hybrid constructions combining traditional and modern techniques also performed poorly. Confined masonry and reinforced concrete buildings generally performed better, although deficiencies in detailing and irregular configurations were frequently observed. Cultural heritage assets sustained predominantly moderate damage, and stabilisation efforts are ongoing. By combining remote sensing and systematic engineering field observations within a common geospatial framework, the study presents a multi-layer reconnaissance framework for interpreting landslide inventories, building damage, and cultural heritage observations together in data-scarce mountainous regions, while highlighting priorities for future research and practice.
The catastrophic landslide and flood events that affected Ischia Island on November 26th, 2022, taking place in the same area previously damaged by an earthquake occurred five years before, underscored the urgent need for integrated multi-hazard risk mitigation strategies in complex volcanic and highly urbanized environments. This study presents a comprehensive Mitigation Measures Plan (MMP) developed under the coordination of the Government Commission for the Emergency and Reconstruction for Ischia Island (GCER) to support land planning, accounting for multiple hazards including rainfall-induced debris avalanches, debris flows, rockfalls, floods, and seismic slope instability at the island scale. Moving beyond conventional single-hazard frameworks, the proposed approach integrates geological, geomorphological, hydrological, and geotechnical data with high-resolution topographic information derived from LiDAR and aerial photogrammetry surveys. Physically based numerical models are adopted to reconstruct hazard scenarios and quantify susceptibility patterns. Rockfall trajectories and runout zones, as well as debris-avalanche and debris-flow propagation, are simulated using 3D software. Flood hazard and inundation dynamics are modelled using two-dimensional hydrodynamic simulations under extreme rainfall scenarios, including climate projections based on intensity–duration–frequency analyses. These datasets are further integrated with historical inventories, seismic microzonation studies, and post-earthquake (2017) and landslide (2022) field inspections to ensure a robust multi-source characterisation of hazard processes. Results highlight strong spatial coupling between steep volcanic slopes, incised drainage networks, and densely urbanized areas, where exposure is significantly amplified. In response, the MMP was developed, combining structural and nature-based remedial measures such as slope stabilization works, debris retention systems, and underground hydraulic diversion tunnels with non-structural strategies including land-use regulation, monitoring systems, emergency planning, and harvesting schedules for protective forests. The Ischia case study demonstrates the effectiveness of integrated multi-hazard approaches in translating advanced numerical modelling into operational risk reduction strategies, providing a transferable framework for other Mediterranean volcanic islands exposed to interacting natural hazards.
Bushfire (wildfire) hazards increasingly threaten buildings in Bushland Urban Interface zones, but property level risk assessment is difficult due to complex building-fire interactions, limited experimental data, and a lack of comprehensive, component specific assessment frameworks. Recent advances in Artificial Intelligence (AI) and remote sensing have substantially enhanced the automated detection and characterisation of Building Envelope Components (BECs).These BECs are key determinants that often increase the likelihood and severity of bushfire damage. Nevertheless, the integration of these detection capabilities with robust, quantitative heat transfer modelling of BECs remains limited. This study proposes an AI-based framework to support bushfire risk assessment that couples BEC detection with physics guided estimation of the maximum rear (non fire-exposed) surface temperature (Tmax) of the fire-exposed outer layer as a key thermal-response variable. Multiple DL object detection models, including YOLO variants and the prompt-conditioned foundation model for object detection, Grounding DINO, were trained and fine-tuned using Google Street View imagery from Queensland, Australia, to detect residential BECs. In addition, multiple Machine Learning (ML) models were trained and evaluated to identify the best approach for estimating Tmax for the detected BECs under bushfire conditions. The Tmax models were trained on experimental data from controlled fire tests on residential wall and window materials conducted by the Wind and Fire Laboratory, Queensland University of Technology, and on simulated data using Fire Dynamics Simulator for Flame Zone exposure curves. The DL-based BEC detection model achieved an mAP50 of 0.78, and the optimal ML model for estimating Tmax achieved a minimum RMSE of 2 °C for walls and windows. The proposed framework facilitates scalable, component-level bushfire vulnerability assessment by integrating detection capabilities with heat transfer modelling. A comprehensive evaluation would require systematic comparison of the estimated Tmax values with established temperature thresholds for each BEC.
Urban metro systems (UMS) are highly susceptible to flood hazards, yet efficient and interpretable post-flood recovery scheduling for large-scale metro networks remains challenging. This challenge stems from coordinating sequential physical restoration across a large action space and distinguishing its effects on multiple service outcomes under time-varying origin-destination (OD) demand, transfer-dependent routing, and above-/underground disruption heterogeneity. Accordingly, this study proposes a service-oriented multi-task deep reinforcement learning (DRL) framework for strategic UMS recovery scheduling. The framework (1) models physical UMS network disruption and recovery, (2) evaluates major service-loss dimensions through network efficiency, operational flow, and ticket revenue, and (3) embeds these dimensions as related but distinct recovery targets. Its shared-trunk architecture learns common recovery representations while retaining service-specific preferences, supporting computationally efficient and interpretable scheduling. Evaluated on the Shanghai Metro under 10-, 100-, and 1000-year flood scenarios, PPO with a learning rate of 1e-4 achieved the best training performance. The learned policy demonstrates cross-scenario robustness, identifies spatially concentrated service-critical domains, reveals the increasing marginal value of additional emergency teams with disruption severity, and reduces cumulative service losses by 0.8%–6.0% relative to the tested heuristic, MILP, and NSGA-II baselines. Overall, this study advances strategic service-oriented UMS recovery scheduling and offers a scalable framework for broader infrastructure systems.
As climate change increases both hurricane risk and damages, there is growing interest in how best to persuade the public to prepare and protect themselves. To increase hurricane preparedness amongst the public, policymakers and emergency management organisations typically issue advice and information. However, novel and creative techniques such as music—where preparedness advice is embedded in lyrics— are also being trialled to address barriers to action like the lack of trust and low efficacy. There is little causal evidence about the behavioural and psychological impact of music on hurricane preparedness, although such evidence can guide the design of effective communication strategies. We present results from an online randomised controlled experiment, conducted in partnership with the Caribbean Disaster Emergency Management Agency and the World Bank, to evaluate the impact of a music campaign to increase hurricane preparedness amongst participants from 11 Caribbean countries (n = 974). We examined impacts on three psychological antecedents of hurricane preparedness, namely, risk perception, self-efficacy and trust in the emergency agency's messages, as well as positive and negative emotions. In the study, participants were randomly exposed to either an unrelated control condition, or text-based information message, or an informational music-based video message urging people to prepare for hurricane season. We found that compared to the control condition, hurricane preparedness intentions were higher amongst participants in the text-based information condition. But the informational music-based video additionally increased self-efficacy and trust in the message, and positive emotions such as pride and happiness, whereas the text-based information message did not. A multiple mediation analysis shows that self-efficacy and trust mediated the positive effect of the music-based video message on intentions to prepare for hurricane season. Our findings suggest that a combination of both text and music-based video messages are valuable for preparing at-risk populations during hurricane season, given that each type of intervention may have distinct behavioural and psychological impacts. Specifically, music might be particularly useful in regions where residents have relatively low self-efficacy and limited trust in hurricane warning messages from disaster management agencies.
The 2023 earthquake in south-eastern Türkiye and north-western Syria was one of the deadliest disasters of the 21st century, with Türkiye being hit particularly hard. In addition to disaster management professionals, spontaneous volunteers played a crucial role in the relief efforts. Through 34 semi-structured interviews, this study explores the factors that enable spontaneous volunteers to participate in relief efforts, the structures that emerge during their self-organisation, and their interactions with formal disaster relief structures.The research identifies motivation, resources, knowledge, needs assessment and disaster memory as key enablers of volunteer-led relief efforts. By using and adapting pre-existing structures, volunteers engaged in creative relief efforts that met immediate needs and supported resilience. Whilst their activities often occurred in parallel with formal disaster response, their co-ordination and integration however were limited. Challenges such as inadequate planning, legal ambiguities and a lack of mutual trust and recognition hampered co-operation between informal and formal responders. Nevertheless, the findings show a strong interest on both sides to work together more effectively.By examining the enablers, organisational structures and interactions of spontaneous volunteers with formal actors, this research provides insights into the role of different forms of agency in crisis response. It contributes to ongoing discussions on how formal disaster risk management institutions can work together with informal volunteer networks to make crisis response systems more equitable, inclusive, and sustainable. The study highlights the need for policies that facilitate co-operation between formal and informal actors, build trust and create mechanisms for better co-ordination in future crises.
Advances in artificial intelligence (AI) have created new possibilities in disaster risk communication, including the automated generation and personalization of emergency alerts. In this paper, we examine how the public perceives AI-personalized alerts compared to standard, human-authored alerts across six dimensions: clarity, trust, relevance, influence, confidence, and certainty. In a randomized controlled trial (RCT), 1158 participants were recruited and assigned to receive either a human-authored alert (Condition 1) or AI-personalized alert (Condition 2) describing a hypothetical disaster scenario, of whom 503 were included in the final analysis. Our results show that different demographic groups rate standard and AI-personalized alerts differently, with moderate variations observed across characteristics such as gender and education levels. To further examine underlying patterns, we use multiclass random forest models to identify predictors of perceived alert quality. We observe differences in the types of features associated with perceptions across conditions. Several factors, including age, gender, and prior disaster experience, are found to be relevant in both conditions. However, for AI-personalized alerts, features related to personal context, such as having dependents and higher education levels (Master's or above), emerge as relatively more influential compared to standard alerts. Overall, our findings suggest that AI-personalized alerts do not consistently outperform standard alerts in perceived quality, but they may shift how individuals interpret and evaluate risk information. These results highlight both the potential and the limitations of AI-based personalization in emergency communication, emphasizing the need for careful design, transparency, and further validation in real-world settings.
Background The Finnish Red Cross has delivered humanitarian assistance in the form of Emergency Response Units to global disasters for several decades. However, the potential use of Emergency Response Units in the Finnish context has not previously been studied. The purpose of this study was, therefore, to explore the cooperation needs between the Finnish Red Cross and public authorities responsible for domestic crisis management, with specific reference to the potential domestic use of Emergency Response Units. Methods This qualitative interview study targeted experts working in crisis preparedness and response within the Finnish Red Cross and public sector organizations (n = 15). The data were analyzed using inductive content analysis. Results The findings answered the research question: How should cooperation between the Finnish Red Cross and the organizations responsible for crisis management in Finland be developed, considering the national utilization possibilities of the Red Cross Emergency Response Units? The results were organized into five categories: Planning and Coordination of Cooperation; Implementation of Joint Activities; Utilization of Human Resources; Utilization of Competence Capital; and Operational deployment contexts and identified needs for Emergency Response Units. Conclusions Cooperation between organizations should be developed through strategic alignment, which requires strong national coordination. Although certain challenges related to operational alignment were identified, cooperation was seen as offering significant potential for strengthening national crisis preparedness and response in evolving environments.
