Reconstructing Disaster Contingency Planning as an Adaptive Policy Instrument: Evidence from Multi-Hazard Risk Governance in West Sumatra, Indonesia

Effective disaster-risk reduction requires the integration of government institutions across relevant policy sectors. In West Sumatra, policy design based on the Sendai Framework has been identified as an important foundation for strengthening disaster-risk reduction, while comparative evidence indicates that governance arrangements and policy design significantly shape post-disaster outcomes. Moreover, post-disaster landscapes may involve contested processes in which waste, value, and institutional responsibilities are redefined, demonstrating that disaster consequences extend beyond immediate physical damage [1-4]. Societal perceptions also influence how calamitous events are interpreted and addressed. Therefore, contingency management must anticipate uncertainty and low-probability “wild-card” events, translate risk knowledge into implementable policy arrangements, and account for interactions among multiple hazards that may produce compound impacts on exposed communities [5-8].

At the institutional level, local disaster-management performance depends on the policy capacity of responsible organizations. Disasters cannot be attributed to hazards or climatic conditions alone, as their consequences are also shaped by exposure, vulnerability, institutional arrangements, and policy choices. Strong institutional design, a multidimensional understanding of resilience, and integrated approaches to interdependent risks are therefore essential components of effective disaster-risk governance [9-13]. Research on the implementation of the Sendai Framework further highlights the importance of connecting disaster-risk reduction with governance and sustainable development, with its relevance also examined in satellite security, tsunami-evacuation infrastructure in Indonesia, and health-system preparedness [14-17]. At the local level, translating the complexity of disaster resilience into practical action additionally requires effective engagement and communication with community leaders [18]. Collectively, this literature supports a Sendai-oriented approach centred on risk knowledge, institutional governance, resilience investment, operational preparedness, and community participation.

This orientation is particularly relevant to Indonesia, where disaster resilience is shaped by local coping capacity, multi-hazard interactions, national risk profiles, and inter-sectoral preparedness for cascading risks [19-22]. In 2024, disasters caused 540 deaths, affected or displaced more than 8.1 million people, and damaged housing, educational facilities, health facilities, and critical transport infrastructure [23]. These impacts demonstrate the need for preparedness policies capable of translating risk information into coordinated and operational action.

Contingency planning is one such policy instrument. Under National Disaster Management Agency Regulation Number 2 of 2023, a contingency plan forms part of disaster emergency planning and may be converted into an emergency operations plan when an emergency occurs [24]. Its operational value depends on whether updated risk data are translated into realistic scenarios, actor responsibilities, command arrangements, resource requirements, evacuation procedures, protection measures, and mechanisms for testing and revision [7, 24-29]. A plan that merely lists hazards and institutions, without specifying activation procedures and measurable resource commitments, may satisfy administrative requirements but remain ineffective during an actual emergency.

West Sumatra provides a critical setting for evaluating this problem because its coastal, urban, river-basin, volcanic, and mountainous areas are exposed to interacting hazards, including earthquakes, tsunamis, floods, flash floods, landslides, eruptions, and cold-lava flows [30]. The province recorded a Disaster Risk Index score of 142.55 in 2024 [31]. However, the central policy concern is not simply the level of risk, but whether changing risk conditions are consistently incorporated into contingency-plan design. Multi-hazard events may simultaneously disrupt transportation, health services, communications, logistics, evacuation, and basic services across administrative boundaries [32-38]. Consequently, contingency planning requires an operational policy system that connects risk evidence, actor responsibilities, available resources, community knowledge, vulnerable-group protection, plan testing, and institutional learning.

Another weakness that needs to be emphasized is the lack of plan testing. Untested contingency plans do not provide assurance that the procedures in the document can be implemented. Without plan testing, weaknesses are only discovered after a disaster has occurred, when corrections become very costly due to the risk to life [39, 40].

Cross-actor coordination is equally crucial, and major disasters cannot be managed by a single institution [41, 42]. While the Regional Disaster Management Agency (RDMA) does have a coordinating function, saving lives, providing health services, providing social protection, transporting communications, providing logistics, providing security, providing evacuations, managing volunteers, and providing initial recovery involve numerous actors. Fragmented coordination will lead to duplication of assistance, delayed evacuations, overlapping authority, weak data collection, and wasted resources [43, 44].

Community empowerment is a crucial dimension in contingency planning, particularly for coastal communities and disaster-prone areas. Communities are the first to experience the impacts of disasters and are often the first responders before formal assistance arrives [45-47]. Accordingly, contingency plans must be developed with meaningful community participation, and communities need to be involved in risk mapping, identifying evacuation routes, determining assembly points, collecting data on vulnerable groups, testing procedures, and evaluating drills.

Table 1. Comparison with the authors’ related studies

From a public policy perspective, contingency planning can be understood as a policy instrument with objectives, targets, actors, resources, procedures, and evaluation mechanisms [48, 49]. Evaluative analysis is relevant because it can assess whether the contingency plan has fulfilled its policy function [50-52].

Table 1 demonstrates that the present study is related to, but analytically distinct from, the authors’ previous publications. The study [7] was the closest antecedent because it evaluated the effectiveness and efficiency of the West Sumatra tsunami contingency plan. However, the present study examines multi-hazard contingency planning across Padang City, Payakumbuh City, Sijunjung Regency, and Lima Puluh Kota Regency; applies six evaluation dimensions; and integrates documentary, interview, observational, and focus group discussion (FGD) evidence into six reconstructive components. The studies [2, 9, 11] addressed broader Sendai-based policy design, policy capacity, and institutional governance, respectively, rather than the operational reconstruction of contingency planning.

Existing studies often examine contingency planning through regulatory compliance, document completeness, or isolated aspects of preparedness. Limited attention has been given to how risk data, multi-hazard scenarios, actor roles, resource gaps, community participation, vulnerable-group protection, testing, and updating interact within a single operational policy system. This study addresses that gap by evaluating contingency planning through six policy dimensions and developing an integrated reconstruction model. Its contribution lies in repositioning contingency planning from a static administrative document into an adaptive, evidence-based, inclusive, and operational policy instrument for multi-hazard disaster governance.

The policy reconstruction referred to in this research goes beyond simply revising the wording of contingency plan documents. Reconstruction must be understood as an effort to redesign policy logic to better align with the nature of risks, regional capacities, community needs, and principles of disaster risk governance. Reconstruction includes updating risk databases, developing multi-hazard scenarios, analyzing resource gaps, refining command structures, integrating local actors, implementing plan testing mechanisms, establishing regular evaluation cycles, and strengthening responsiveness to vulnerable groups. In this way, contingency plans can move from being mere compliance documents to operational policy instruments.

Accordingly, this study addresses two research questions: (1) To what extent does the existing design of disaster contingency planning in West Sumatra satisfy the dimensions of effectiveness, efficiency, adequacy, appropriateness, equity, and responsiveness? (2) How can the identified policy-design weaknesses be reconstructed into an adaptive and operational contingency-planning model for multi-hazard governance?

The NVivo analysis in this study was constructed through a coding structure that stemmed from the research objectives, namely analyzing the existing design of contingency planning and formulating a policy reconstruction based on evaluative analysis. The initial coding structure used six dimensions of policy evaluation: effectiveness, efficiency, adequacy, appropriateness, equity, and responsiveness. These six dimensions were then developed into empirical sub-descriptors relevant to contingency planning, such as risk databases, multi-hazard scenarios, actor clarity, command structures, resource mapping, plan testing, document updates, community participation, and protection of vulnerable groups.

In NVivo, each instrument is positioned as a data source with a distinct analytical function. Policy documents were used to assess the formal design, normative standards, and consistency of the plan's substance. In-depth interviews were used to capture actors' perceptions, experiences, obstacles, and assessments. Field observations were used to verify the condition of facilities, evacuation routes, assembly points, equipment, shelters, communications, and command centers. FGDs were used to validate findings, compare actors' perspectives, and test the feasibility of reconstruction alternatives.

The coding results indicate that contingency planning issues cannot be interpreted as a single issue. Each evaluation dimension is interconnected. Weaknesses in the risk database impact scenario quality. Weaknesses in scenario analysis impact resource requirements calculations. Weaknesses in resource analysis impact response adequacy. Weaknesses in coordination impact activation effectiveness. Weaknesses in updating impact policy responsiveness. Weaknesses in inclusiveness impact protection alignment. Consequently, NVivo results not only show a list of descriptors but also the relationships between them that shape the structure of the policy problem.

Frequency of coding references is used to help identify issue intensity but should not be treated as a sole measure. High-frequency descriptors indicate that the issue appears frequently in the data, while lower-frequency descriptors can still be considered strategic if they have significant consequences for public safety. For example, protecting vulnerable groups may not always be the most dominant descriptor in documents, but theoretically and practically, it is a crucial component of contingency planning. Accordingly, interpretation of results is conducted through a combination of coding references, matrix coding, analytical memos, and inter-source triangulation.

The final NVivo structure contained seven principal empirical nodes and six evaluative parent nodes. Coding frequencies were used to identify thematic prominence, while triangulation determined analytical significance. Documents mainly emphasized plan substance and risk data; interviews highlighted coordination and resources; observations verified facilities and operational capacity; and FGDs linked the identified weaknesses to reconstruction alternatives.

3.1 Results of policy document analysis

Document analysis indicates that contingency planning has a relatively clear regulatory basis and institutional format. Policy documents generally contain information on the type of threat, vulnerable areas, involved actors, coordination structures, and some resource requirements. The evidence suggests that administratively, contingency plans have been positioned as part of the regional preparedness system. However, coding results also indicate that administrative completeness has not been fully followed by operational depth. The coding generated 18 references to contingency-plan substance, 17 to risk-data foundations, 16 to multi-hazard scenarios, 15 to actor-role allocation, 14 to resource gaps, 13 to testing and updating, and 12 to vulnerable-group protection. Several documents identified hazards and responsible institutions but did not specify activation time, command channels, resource commitments, or reporting procedures. Other documents presented resource inventories without comparing available capacity with scenario-based requirements.

The most prominent descriptor in the document analysis is the substance of the contingency plan. This sub-descriptor includes clarity of objectives, completeness of components, consistency between risk analysis and action plans, and the document's alignment with policy guidelines. Documents tend to be strong in general descriptions, but not always strong in linking risks to measurable actions. Some documents mention the threat of earthquakes, tsunamis, floods, landslides, eruptions, or lahars, but operational scenarios still need clarification, particularly regarding the number of affected populations, shelter capacity, health service needs, and logistical requirements.

The second descriptor is the risk database. Risk data has been used as background, but has not always been the basis for calculating needs. Within an evidence-based policy framework, risk data should be used to prioritize areas, determine worst-case scenarios, calculate resources, and develop evacuation mechanisms. If data is used only as an initial narrative, risk plans lack sufficient precision. Document findings indicate that the use of risk data still needs to move from a descriptive to an operational function.

The third descriptor is the multi-hazard scenario. Documents generally list the main threats but do not fully integrate multiple and cascading risks. However, in the context of West Sumatra, threats can be interconnected. Eruptions can trigger cold lava flows during heavy rainfall. Earthquakes can trigger tsunamis, damage infrastructure, disrupt communications, and hinder evacuation. Floods can disrupt health services and logistics distribution. The limitations of these scenarios indicate the need for reconstruction to better align plans with actual risk dynamics.

The fourth descriptor is the division of actor roles. Documents list numerous institutions but do not always detail the division of tasks. Many formulations are general, such as coordinating, supporting, or implementing according to authority. Such formulations are difficult to use in an emergency because they do not answer the questions of who does what, when, with what resources, through which chain of command, and to whom to report. Thus, the division of roles needs to be translated into testable operational tasks.

The fifth descriptor is the resource gap. Documents generally contain an initial inventory, but it doesn't always compare available capacity with scenario requirements. A resource list is insufficient without information on condition, location, actual capacity, mobilization time, owner, and external support mechanisms. This finding suggests that the inventory needs to be developed into a resource gap analysis.

The sixth descriptor is testing and updating. Documents acknowledge the importance of simulations and regular updates, but follow-up mechanisms are not always clear. Plan testing schedules, success indicators, those responsible for evaluation, and revision procedures need to be strengthened. Without a learning cycle, contingency plans risk becoming static documents that no longer align with the latest risks and capacities. The seventh descriptor is the protection of vulnerable groups. Vulnerable groups are frequently mentioned in documents, but this is not always translated into operational data, specific evacuation procedures, shelter services, and accessible communications. This situation suggests that the equity dimension still needs to be strengthened in plan design.

Figure 1 shows that a contingency plan document cannot be judged solely by the presence of its components. The quality of the document must be seen in the relationships between the components. If risk data is not used to determine scenarios and resource requirements, the document has the potential to be administratively strong but operationally weak. The strongest document-based themes concerned plan substance, risk data, and scenario development. However, the relatively high references to role allocation, resource gaps, testing, and vulnerable-group protection indicate that formal completeness was not consistently supported by operational detail.

tu_pian_17.png

Figure 1. Coding matrix in policy document analysis

Document analysis showed that several contingency plans identified hazards, institutions, and resources but did not consistently translate them into measurable operational arrangements. For example, one document stated that relevant agencies should coordinate according to their respective responsibilities, without specifying activation time, command channels, resource commitments, or reporting procedures. This evidence supports the finding that documentary completeness has not yet ensured operational clarity.

3.2 In-depth interview results

In-depth interviews revealed a more complex dynamic than document analysis. Government informants generally considered contingency plans to be a crucial part of preparedness. However, informants from Non-Governmental Organizations, volunteers, communities, and vulnerable groups highlighted the gap between documents and practice. This difference in perspective is important because it demonstrates that preparedness cannot be assessed solely from the perspective of policy makers, but also from the perspective of implementing actors and the communities who will use the plan in an emergency.

The dominant descriptor in the interviews was cross-actor coordination. Informants stated that coordination is a determining factor in plan activation. Although coordination structures are in place, practice still relies heavily on personal relationships, field experience, and informal communication. Informal communication often helps expedite responses, but also poses sustainability risks if personnel change or communication networks are disrupted. These findings indicate that coordination needs to be institutionalized through clear procedures, proven communication channels, and a clear division of roles that all actors understand.

The second descriptor is resource capacity. Informants assessed that available capacity may be sufficient for specific incidents, but not necessarily for large-scale scenarios or trans-regional disasters. The issue is not simply the quantity of resources, but also their readiness, location, accessibility, and mobilization time. Registered equipment may not be ready for use. Registered shelters may not meet basic needs. Available vehicles may not be able to reach affected locations if roads are blocked. Thus, capacity needs to be assessed in actual terms, not nominally.

The third descriptor is planning testing. Informants acknowledged that simulations and exercises have been conducted on several occasions, but not all actors have consistently participated. Simulations have also not always been followed by evaluations that impact document revisions. In some cases, exercises have emphasized ceremonial aspects rather than critical testing of scenarios, commands, resources, and protection of vulnerable groups. These findings suggest that plan testing should be positioned as a policy evaluation mechanism, not simply a socialization activity.

The fourth descriptor is community participation. Communities are often positioned as recipients of information, rather than actors in plan development. Yet, they possess knowledge of local routes, safe points, vulnerable groups, and community resources. Community informants emphasized that limited outreach has left many residents unaware of the contents of contingency plans. Community engagement needs to be expanded beyond outreach to participatory mapping, community simulations, and evacuation route evaluations.

The fifth descriptor is the use of risk data. Several informants assessed that risk data is available, but its updating and use in decision-making are still suboptimal. Data on population, vulnerable groups, shelters, evacuation routes, and resources need to be synchronized across agencies. Data inconsistencies have the potential to cause confusion during emergency response.

The sixth descriptor is the inclusiveness of services. Informants from vulnerable groups assessed that evacuation and shelters do not fully address special needs. Physical accessibility, easy-to-understand information, the need for medication, assistance, and sanitation still need to be strengthened. This finding indicates that the dimensions of equity and responsiveness must be core components of policy reconstruction.

Figure 2 shows that the interviews yielded varying perspectives. These differences are not data weaknesses, but sources of triangulation. A good contingency policy must be able to answer the assessments of the government as the drafter, the Regional Apparatus Organizations as sectoral implementers, vertical institutions as technical support, and the community as end users. This indicates that cross-actor coordination and resource capacity were the most prominent concerns in the interviews. This pattern supports the finding that contingency plans are not always understood or activated consistently across institutions, especially when personnel, resources, or risk conditions change. Cross-actor coordination was the most frequent interview theme, with 19 coding references, followed by resource capacity with 18 references. Plan testing generated 16 references, community participation 15 references, the use of risk data 14 references, and service inclusiveness 13 references. These patterns show that operational coordination and actual resource readiness were the most prominent concerns, although participation and inclusive protection remained strategically important.

tu_pian_18.png

Figure 2. Coding matrix in in-depth interviews

3.3 Field observation results

Field observations were used to verify whether the components outlined in the document were actually available and usable. Observation findings revealed variations in preparedness between regions and facilities. Some locations had evacuation route signs, assembly points, shelters, equipment, and communication facilities, but their quality and actual use were uneven.

The first descriptor in the observation was evacuation routes and assembly points. In some areas, evacuation routes were present, but others were obscured by buildings, vegetation, billboards, or were not clearly visible at night. Some evacuation routes passed through narrow roads, busy intersections, or areas potentially disrupted during a disaster. The evidence suggests that evacuation routes need not only be specified in the document but also tested under various conditions, including rain, nighttime, heavy traffic, and public panic.

The second descriptor is shelter and basic facilities. Some locations designated as shelters have sufficient space, but not all are equipped with clean water, sanitation, lighting, accessibility, dedicated service areas, and refugee management mechanisms. Shelters that do not meet basic needs can create new problems during the emergency phase. Consequently, shelter assessments must include service capacity, quality, access, and adequacy.

The third descriptor is equipment readiness. Observations reveal discrepancies between inventory and actual conditions. Some equipment is available but requires maintenance, component replacement, or more strategic placement. This underscores the importance of periodic equipment readiness audits.

The fourth descriptor is command centers. Some command centers have communication and monitoring facilities, but system redundancy is not always robust. Reliance on electricity, internet networks, or specific devices can be a weak point when infrastructure is impacted. Alternative communication and backup procedures need to be incorporated into plans.

The fifth descriptor is emergency communication systems. Public information has been provided in the form of signs, boards, or outreach materials, but public understanding cannot be assumed. Risk communication should be tested through simulations and tailored to target groups. The sixth descriptor is accessibility for vulnerable groups. Observations show that not all routes, shelters, and facilities have adequate access for people with disabilities, the elderly, pregnant women, or residents with special needs. This situation reinforces the need to incorporate universal design and specific procedures into policy reconstruction.

Figure 3 distinguishes nominal capacity recorded in planning documents from actual capacity verified through facility condition, accessibility, maintenance, and operational usability. Field observations generated 18 references to evacuation routes and assembly points, 17 to shelters and basic facilities, 15 to equipment readiness, 14 to command centres, 13 to emergency communication systems, and 12 to accessibility for vulnerable groups. These records confirmed that the availability of facilities in planning documents did not always correspond to their condition, accessibility, maintenance, or operational usability.

tu_pian_19.png

Figure 3. Coding matrix in field observations

3.4 Focus group discussion results

The FGD served as a forum for validation and synthesis. FGD participants generally agreed that contingency plans remain necessary as a coordination framework, but their design needs to be made more operational, simple to use, testable, and easy to update. Participants noted that overly lengthy documents often make them difficult to use in emergency situations. Accordingly, contingency plans need to be supplemented with operational summaries, task cards, contact lists, resource maps, and clear activation procedures.

The first descriptor in the FGD was a dynamic risk baseline. Participants emphasized that plans should utilize the most recent risk data, including changes in settlements, infrastructure, population, vulnerable groups, and service capacity. Data should not be confined to appendices but should form the basis for determining scenarios and resource requirements. The second descriptor was a multi-hazard scenario. Participants assessed that plans need to account for interactions between hazards. In the context of West Sumatra, disasters can evolve from a single threat into a series of crises involving access to transportation, health, logistics, communications, and evacuation. Consequently, scenarios must include cascading impacts and cross-sector response strategies.

The third descriptor is operational governance. Participants emphasized the importance of clear and testable role allocations. Each actor needs to know their tasks, resources, activation times, and reporting mechanisms. The coordination structure must be implemented not only on paper but also in exercises. The fourth descriptor is resource gap analysis. Participants agreed that resource lists need to be updated and verified. Needs in specific scenarios should be compared with actual capacity. If there are gaps, plans should outline strategies for meeting them, whether through support from the province, other districts/cities, businesses, volunteers, or the community.

The fifth descriptor is inclusive protection. Participants stated that vulnerable groups need to be involved in planning and simulations. Disaggregated data, accessible routes, shelter services, dedicated communication, and mentoring should be part of the design. The sixth descriptor is testing and learning. Participants emphasized that simulations should produce evaluation records and document revisions. The FGD recommended after-action reviews as a routine mechanism following drills and disaster events. The seventh descriptor is regular updates. Participants assessed that plans should have a clear review cycle, designated person(s), indicators, and budget support. Updates should not be dependent on temporary projects but should be an ongoing institutional function.

Figure 4 shows how the FGD transformed the identified weaknesses into policy-design priorities. Dynamic risk data, multi-hazard scenarios, operational governance, resource-gap analysis, inclusive protection, testing, and periodic updating emerged as the main requirements for reconstruction. The FGDs indicated that policy reconstruction needs to be geared toward a practical system, not just a lengthy document. Participants emphasized the importance of operational summaries, task cards, resource maps, and user-friendly update mechanisms. FGD coding generated 18 references to dynamic risk data, 17 to multi-hazard scenarios, 17 to operational governance, 16 to resource-gap analysis, 15 to inclusive protection, 15 to testing and institutional learning, and 14 to periodic updating. The need for updated risk data was emphasized by provincial, local-government, and academic participants. Clear role allocation and resource-gap analysis were supported by disaster-management and technical-agency representatives. Inclusive protection was particularly emphasized by community and vulnerable-group representatives.

tu_pian_20.png

Figure 4. Coding matrix in focus group discussion (FGD)

3.5 Integration of contingency planning evaluation dimensions

The integration of data from documents, interviews, observations, and FGDs indicates that contingency planning has strengths in normative recognition and the existence of a basic framework, but still requires strengthening in operational aspects. In the effectiveness dimension, the plan has not yet fully become a working guideline used by all actors. Activation still relies on experience and informal communication. In the efficiency dimension, data and resources are still scattered, potentially leading to duplication and delays.

In the adequacy dimension, available capacity does not always match the needs of major scenarios. Inventories need to be developed into gap analyses. In the appropriateness dimension, plans still tend to be generic and not fully specific to regional variations and risk characteristics. In the equity dimension, protection of vulnerable groups still needs to be translated into operational procedures. In the responsiveness dimension, evaluation and learning have not yet fully resulted in document updates.

The integrated findings indicate that design weaknesses arise from weak relationships between components. Risk data doesn't always influence scenarios. Scenarios don't always form the basis for resource calculations. Resources aren't always linked to actors and commands. Simulations aren't always linked to revisions. Public participation isn't fully connected to formal structures. Thus, policy reconstruction needs to be carried out systematically.

Figure 5 indicates that adequacy, effectiveness, and responsiveness were the most prominent evaluative dimensions. Nevertheless, equity remained strategically important because failures in accessibility and protection directly increased risks for vulnerable groups.

tu_pian_21.png

Figure 5. Coding matrix on evaluation dimension integration

The integration matrix emphasizes that policy reconstruction must address the interrelationships between dimensions. Effectiveness requires scenario appropriateness, resource adequacy, coordination efficiency, protection equity, and responsiveness to reforms. Thus, no single dimension can be improved in isolation from the others.

Figure 6 presents the final analytical synthesis. The model links risk data with scenario development, scenarios with resource requirements, actor roles with command and coordination, and testing with evaluation and updating.

tu_pian_22.png

Figure 6. Coding matrix in policy reconstruction

This finding indicates that disaster contingency planning should not be understood merely as a formal administrative document required by regulation, but as a strategic policy instrument that determines how risk knowledge is translated into collective action before and during emergencies. The case of West Sumatra shows that the existence of contingency planning documents does not automatically guarantee operational preparedness. Although the plans have a relatively clear normative basis and contain several essential components, their policy function remains limited when risk data are not fully transformed into realistic scenarios, actor roles are not translated into operational tasks, resource inventories are not developed into gap analyses, and simulations are not systematically used as learning mechanisms.

The integrated findings from document analysis, interviews, field observations, and FGD indicate that the main weakness of contingency planning lies in the disconnection among policy components. Risk data often function as descriptive background rather than as the basis for scenario formulation and resource calculation. Institutional coordination is formally recognized but is not always supported by clear activation procedures, communication flows, task cards, and accountability mechanisms. Resource capacity is frequently listed in documents, yet its actual readiness, location, accessibility, condition, and mobilization time are not consistently verified. Similarly, vulnerable groups are acknowledged in planning narratives, but their specific needs are not always translated into accessible evacuation routes, inclusive shelters, disaggregated data, tailored communication, and assisted evacuation procedures.

Theoretically, this study contributes to the literature on disaster risk governance by integrating policy evaluation, adaptive governance, and contingency planning into a single analytical framework. It shows that effectiveness alone is insufficient for assessing disaster preparedness policies; plans must also be efficient, adequate, appropriate, equitable, and responsive. Practically, the reconstruction model provides local governments with a usable framework for transforming contingency plans from static compliance documents into adaptive preparedness systems. By strengthening the linkage between evidence, actors, resources, community participation, and institutional learning, contingency planning can better support rapid response, reduce disaster losses, protect vulnerable groups, and enhance regional resilience in multi-hazard contexts.