A GeoAI-Driven and Decision-Oriented Methodology for Multi-Hazard Early Warning System Development

The intensity and frequency of natural hazards have increased unprecedentedly, resulting in devastating impacts on lives and economies. The present challenge for practitioners and policymakers is to reduce such impacts and related risks with innovative measures. A multi-hazard early warning system has been recognized as a crucial element in most disaster risk reduction measures. However, recent hazard incidents revealed that existing early warning systems are not completely able to save lives and reduce economic losses due to many reasons. Hence, this paper is conducted to provide a comprehensive understanding of the enablers to be considered when developing a fully comprehensive multi-hazard early warning system. Further, the study presents the benefits of establishing an effective multi-hazard early warning over single early warning systems. The paper is based on a narrative review applying a systematic basis for selecting research papers for the study. According to study results, three aspects to be considered when developing and operating an effective multi-hazard early warning system. Policy, legislative and institutional arrangements; social and cultural considerations and technological and scientific arrangements are the three categories. Policy, legislative and institutional arrangements contain governance, political recognition, mainstreaming early warning into development planning, stakeholder partnerships, periodic feedback, empowerment of local authorities and provision of resources/infrastructure. Social and cultural consideration includes training and capacity building, awareness and education, planning and preparedness, community engagement and empowerment and consideration of gender perspective and cultural diversity. Finally, integration of technological and scientific knowledge, risk information, hazard warning dissemination and communication, and monitoring and forecasting are the technological and scientific arrangements to be assured for an effective multi-hazard early warning for disaster risk reduction and enhancing resilience.

Indonesia is a disaster-prone region facing multiple hazards. The impact has cost human casualties, injuries, damages to assets, environment and regional development. In order to protect the people and assets, the government has increased the capacity to face and anticipate disaster which is through a multi-hazard early warning system (MHEWS). The academic sector in Indonesia, especially prominent universities such as Institut Teknologi Bandung and Universitas Gajah Mada has contributed positively to the development of MHEWS and the Land Slide Early Warning System. This research aims to examine the role of universities in Indonesia related to the establishment of an MHEWS in Indonesia by answering what are the roles and the challenges. The research methodology is through a desk study and document review of government and organizations reports using qualitative and descriptive analysis. The findings of the research indicate the role of universities in the MHEWS is in line with the concept of tri dharma of universities which research, higher education and outreach. This means the role of universities to conduct research related to early warning systems, higher education through having program studies related to disaster management and outreach the role to train the community, organizations, and government through professional networks. Unfortunately, the involvement of the eastern part of Indonesian universities still needs to increase capacity, awareness, and access in utilizing the MHEWS at the local level.

In recent years, great progress has been made in the development of knowledge and practices related to the multi-hazard early warnings and in strengthening the related multi-stakeholder partnership. Global initiatives are gaining momentum to improve multi-hazard early warning systems (MHEWS) and so boost the resilience of the most vulnerable countries to extreme weather/water /geological, environment and health-related events and its impacts to the sustainability of social and economic development. However, understanding its interconnectivity and interoperability and developing impact-based and risk-informed methodologies in its integration in the all relevant hazard aspects are still needed to be strengthened. This article presents an overview of advances and challenges in multi-stakeholder partnership for improving MHEWS in an integrated way in order to better achieve the Target-G of the Sendai Framework on DRR, Paris Agreement on Climate Change and SDGs. We focus on major international cooperation and partnerships on MHEWS and its applications, but not limited to, the International Multi-Hazard Early Warning System Network (IN-MHEWS), Climate Risk and Early Warning Systems (CREWS), Regional Integrated MHEWS in Africa and Asia (RIMES); and many specific MHEWSs interfaces, such as Global Disaster Alerting Coordination System (GDACS), Global Multi-hazard Alert System (GMAS), MeteoAlarm/Alert Systems, All Risk Integrated System TOwards Trans-boundary hoListic Early-warning—European Natural Hazards Scientific Partnership (ARISTOTLE-ENHSP), etc. as well as relevant multi-stakeholder partnership platforms to connect related MHEWS with its application networks between international organizations, governments, non-governmental organizations (NGOs), private sectors at regional and national levels, such as United Nations Disaster Assessment and Coordination (UNDAC), Global Water Partnership(GWP), Environment and Humanitarian Action (EHA) Network, Global network on Monitoring, Analysis, and Prediction of Air Quality (MAP-AQ), Public Health Emergency Operations Centre Network (EOC-NET), Humanitarian Networks and Partnerships Week (HNPW), and Forecast-based Financing mechanism and programme (FbF) etc. The further actions to facilitate UN Member States for improving its capacity in MHEWS partnership at national and local levels through strengthening multi-stakeholder partnership at international and regional levels are discussed and recommended in the conclusion and discussion.KeywordsImpact-based forecasting and integrated risk managementInteractivityInteroperabilityMulti-hazard early warning systemsMulti-stakeholder partnerships governanceRisk interconnectivityPublic and private partnership (PPP)And public and public partnership (PUP)

Recent international agreements such as the Sendai Framework for Disaster Risk Reduction 2015–2030, the 2030 Agenda for Sustainable Development and the Paris Agreement have all recognized the importance of developing and operationalising multi-hazard early warning systems that integrate the specificities of single-hazard early warning systems in a holistic, systematic and coordinated manner to promote synergies and maximize efficiency. While much progress has been made in recent years towards the advancement of knowledge and practice related to early warning systems worldwide, the lack of multi-disciplinary and transboundary cooperation among and across communities of scientists, decision-makers and practitioners continues to be a key challenge for the successful establishment and operation of these systems. To address this gap, major international and national organizations have collaborated to establish the International Network for Multi-Hazard Early Warning Systems (IN-MHEWS), with the aim of facilitating knowledge sharing and capacity development for multi-hazard early warning systems around the globe. This paper presents an overview of advances and challenges in promoting a multi-hazard and systematic approach to early warning, as well as the aim, objectives and expected contributions of this newly established Network.

The Africa Multi-Hazard Early Warning and Action System for Disaster Risk Reduction (AMHEWAS for DRR) is a joint effort, led by the African Union Commission (AUC) in coordination with Regional Economic Communities and Member States and with the technical and scientific support of UNDRR and CIMA Foundation, aimed at strengthening Africa's resilience to natural hazards. This comprehensive system encompasses a multi-scale approach, spanning from continental to regional and national levels, to enhance early warning capabilities and promote effective disaster risk management strategies. On the continental scale, AMHEWAS operates through a network of Situation Rooms. These interconnected hubs facilitate real-time information exchange, coordination of response efforts, and dissemination of advisories on potential threads and related impacts to national institutions. To ensure standardized operational procedures across the continent, AMHEWAS has established unified standard operating procedures, ensuring consistent application protocols and methodologies. Central to AMHEWAS' approach is the Continental Watch (CW), an impact-based forecast bulletin for rain, wind and flood hazards, that synthesizes insights from automated impact-based forecast systems. The CW provides timely and actionable information to decision-makers across the continent, enabling proactive measures to mitigate potential disaster impacts. Ongoing disasters can trigger Disaster Situation Reports (DSRs), co-produced by the AUC with the affected Regional Economic Communities (RECs) and the national AMHEWAS stakeholders, for informing disaster risk reduction (DRR) efforts and ensuring timely and appropriate responses to emergencies.  AMHEWAS integrates risk data and forecasting products from global and regional authoritative sources to produce advisories as a combination of hazards, exposure, vulnerability and national copying capacity. Based on the possible expected impacts in the next 5 days, advisories are issued with a threshold-based mechanism with 4 levels of activation of the system. High level is related with the potential of the estimated impacts to overcome the capacity of the countries, while for lower advisories the effects are expected to be managed by national or subnational authorities. The potential impacts are estimated with an innovative automatic approach, that involves the overlap of the forecasted hazards, with layers of exposed elements, taking into consideration the lack of copying capacity derived from the INFORM database. In order to maximize the robustness of the forecasts AMHEWAS adopts a multimodel approach. As regards wind and rain, the forecast is carried out considering the combination of different meteorological global models. As regards flood, the reference model is GLOFAS, combined for the Great Horn of Africa region with the results of the impact-based flood forecast system FloodPROOFS East Africa (FPEA). FPEA is an operational system based on open-source technologies that employs an impact-based approach, integrating weather forecasting, hydrology and hydraulic modeling, as well as risk assessment to provide accurate and actionable flood forecasts up to five days in advance. Given its cross-border nature, the system allows for a comprehensive approach to large-scale hydrological assessment, easily scalable in an operational framework on a national scale. AMHEWAS is working on further integration of regional forecasting products from WMO specialized centers and national level, in order to improve the risk knowledge and information products generated.

In our increasingly interconnected world, the compound, cascading, and systemic nature of risks takes on a new dimension and meaning. Africa is one of the continents predicted to be most adversely affected by climate change due to high levels of exposure to increasingly intense and frequent hazards and their interactions with intrinsic drivers of vulnerability, resulting in compounding and complex risk profiles. For instance, Africa is increasingly subject to droughts, floods, heatwaves, cyclones and other climate-related hazards. State-of-the-art people-centered, impact-based Multi-Hazard Early Warning Systems (MHEWS) and Anticipatory Action (AA) have to address this growing complexity. While these approaches are widely recognised, implementation of such systems at different scales, from local, national to regional levels, is still impeded by gaps in knowledge and capacities to act on such knowledge. This includes, for instance, how certain hazards can affect the people and property to be protected (impact-based forecasting), or how to generate and communicate user-focused risk-information and knowledge that enables equitable and gender-responsive risk reduction, and how to define triggers and coordination structures that enable early release of funds in anticipation of impact. This paper will introduce to the Network of Centres of Excellence (NoE) which was initiated by the UN Office for Disaster Risk Reduction (UNDRR) in order to contribute to the African-Union led African Multi-Hazard Early Warning and Action System (AMHEWAS) initiative. The aim of the NoE is to enhance the general awareness regarding multi-hazard risks and to strengthen stakeholders’ capacity across national, regional, and continental scales to collect, exchange and analyse risk data and knowledge to adequately inform multi-hazard and impact-based early warning systems and early actions across Africa. The NoE also envisions an integrative, participatory Capacity Development Framework, that incorporates African Indigenous Knowledge Systems, serving towards a general capacity strengthening effort in the region and contributing to the growth of an African-based Community of Practice (CoP) for Disaster Risk Reduction (DRR) and Emergency Response Preparedness (ERP) coupled with an effective Science-Policy-Action feedback loop resulting in tangible interventions on the ground from an African perspective.

This chapter presents an assessment of the effectiveness of the tropical cyclone early warning system (EWS) in Zimbabwe, with a special focus on Cyclone Idai, which affected the country in March 2019. The study used a household questionnaire survey, interviews, field observations and official documents to identify gaps in major components of the EWS value chain, which contributed to the dysfunction of the EWS, resulting in fatalities, loss and damage of property and infrastructure. Lack of risk knowledge, poor instrumentation for accurate forecasting and nowcasting, weak institutional coordination and inadequate resources—that hinder critical government institutions from executing their mandates in the system—and poor early warning communication are some of the identified factors that contributed to the dysfunction of the EWS. The study proposes a set of enabling conditions, as corrective measures to improve all the EWS components. The private sector is an underutilised actor in the EWS, and we recommend that, rather than engaging the private sector in the post-disaster response, the private sector needs to be incorporated in pre-disaster EWS activities. A people-centred, impact-based and multi-hazard EWS, which is founded on shared resources, risk knowledge and expertise, should be considered. Further research may focus on a cost-benefit analysis of the EWS, to provide an evidence base to inform decisions in EWS investment and strategies.KeywordsCyclonesEarly warning systemsIdaiRisk knowledgeForecasting

We present a novel implementation of the impact-based drought monitoring and early warning component of the Africa Multi-Hazard Early Warning and Action System for Disaster Risk Reduction (AMHEWAS for DRR). AMHEWAS is a collaborative initiative led by the African Union Commission (AUC), in partnership with Regional Economic Communities, Member States, and with the technical and scientific expertise of UNDRR and CIMA Research Foundation. Its goal is to enhance Africa's resilience to natural hazards. The system adopts a comprehensive, multi-scale framework, integrating efforts at continental, regional, and national levels to strengthen early warning systems and advance disaster risk management strategies.A key component of AMHEWAS is the Continental Watch (CW), an impact-based bulletin focused on rain, wind, flood, and drought hazards. The CW contains advisories on the levels of potential disaster impacts and is used by the African Union Commission to alert national authorities about potential threats. The CW consolidates data from automated monitoring and forecasting systems, providing decision-makers across Africa with timely, actionable information. This enables proactive interventions to reduce the potential impacts of disasters.The drought bulletin, released to interested parties monthly, monitors drought hazard in near-real time at the continental scale. It uses openly available datasets to evaluate emerging drought conditions at different time scales, to ensure a diverse range of potential impacts is captured in the bulletin. For short-time drought evaluation (1-3 months), the Combined Drought Indicator (CDI) is used as hazard indicator, whereas for longer timescales, a 12-month Standardised Precipitation Index (SPI12) is used.Hazard information from these indicators is paired with tailored layers for exposure and vulnerability to evaluate emerging potential drought impacts in various sectors. Results are presented monthly in the bulletin for the entire African continent and are aggregated at national and sub-national scales.

Coastal hazards taken place all over the world have ended up with major devastations such as loss of life, property damage and environmental degradation more often. Sri Lanka, being an island in the Indian ocean is much vulnerable to coastal hazards; the unprecedented human disaster that caused by the Asian Tsunami was one of the worst examples. Present studies show that there are lot of programs carried out by the government and non-government organizations to drag down the adversity towards the coastal communities in the country who are much vulnerable to coastal disasters, focusing all the segments of basically social, economic and environmental aspects. Evacuation of coastal communities safely in disastrous situations is one of the main requirements within the disaster recovery framework. Having effective multi-hazard early warning (MHEW) systems and procedures in place in the coastal regions of the country helps to reduce the disaster risk. Although there are policies and practices for the evacuation of coastal communities including MHEW systems, the present study shows that the attention and preparedness towards the evacuation of people with special needs in the coastal communities is more or less omitted by the responsible parties. Moreover, when disasters strike, people with special needs meet various discriminations in terms of evacuation to a shelter or relief points and more often are overlooked from the emergency preparedness, response and recovery stages. Therefore, the present research focuses on determining the gaps in the evacuation of people with special needs in the context of MHEW systems during coastal hazards in Sri Lanka. Open interviews and telephone interviews were conducted in few major coastal cities and understood the likelihood of the obstacles, issues and the perception towards the evacuation of people with special needs. A questionnaire was developed according to the open interviews and fine-tuned in three sandpit events employing various local and foreign professionals in academics. The questionnaires were distributed online and the responses were analyzed. Eventually, a substantial number of gaps were identified in supporting the prevailing systems in Sri Lanka and was arrived at conclusions to impact the government through the findings in order to make the evacuation of people with special needs during coastal disasters more efficient.

Use of a Bayesian Network for coastal hazards, impact and disaster risk reduction assessment at a coastal barrier (Ria Formosa, Portugal)

SummaryThis perspective critically examines the challenges and opportunities of implementing people-centered multi-hazard early warning systems (MHEWS) in the Global South. Despite global initiatives, such as the Early Warnings for All initiative, operational realities lag behind. By exploring the needs of the most vulnerable and how core concepts of multi-hazard thinking (e.g., hazard interrelationships and vulnerability dynamics) integrate into different pillars and cross-cutting components of an MHEWS, the perspective highlights a mismatch between current ambitions and realities on the ground. Drawing on extensive experience from Practical Action, we identify opportunities to move toward MHEWS through outlining potential entry points in research, policy, and practice. We emphasize a need for localized, inclusive strategies that genuinely address the needs of the most vulnerable populations and fully encompass the meaning of multi-hazards, including hazard interrelationships, the dynamics of risk components, and the complexity of multi-hazard impacts.

Science, policy, and practice have long accepted that disasters occur due to vulnerabilities rather than hazards, yet approaches to warnings still tend to be hazard-focused. Multi-hazard early warning systems (MHEWS) are meant to provide warnings for many hazards, whether sequential, simultaneous, or cumulative, and even if ostensibly independent. Despite their advantages, MHEWS display the same inherent limitation as most warning systems: they focus on hazards, without sufficient attention given to vulnerabilities. This paper aims to explain and overcome this limitation of hazard-focused warnings and warning systems. Following discussion of the ethos behind, advantages of, and limitations regarding MHEWS including with respect to the United Nation's "Early Warnings for All" initiative, this article proposes a complement to MHEWS: All-Vulnerability Warning Systems (AVWS). The implications of and further work for implementing AVWS are discussed, highlighting the vision that warning systems as social processes should: (1) Across different people, address vulnerabilities conferring widely varying experiences to the same hazard. (2) For the same people, address vulnerabilities conferring similar difficulties to different hazards.

This study addresses the urgency to re-strategize our local action to prevent future climatic risk as a result of extreme weather events, urbanization, anthropogenic activities in a changing climate. While progress has been made in implementing the Sendai Framework for Disaster Risk Reduction since its adoption in 2015, no country is on track to achieve the outcome and goal by 2030. Malaysia is not an exception. A holistic approach to multi-scale disaster risk reduction and climate resilience is critically needed to examine new prospective agenda for accelerated action. This study provides a new insight into galvanizing technological advancement, multi-tier partnership and community-led approach to entail more coordinated and programmatic action towards translating resilience thinking approach into risk-informed decision-making. Equipping cities and communities with knowledge and capabilities to manage complexity of risks is a step forward to re-build a resilient society and rejuvenate resilience thinking. UNDRR’s global reports indicated that by providing a 24-hour early warning can reduce the resulting damage by 30%. Therefore, investing in the development of people-centered, end-to-end, multi-hazard early warning system (EWS) is highly regarded to support the 2027’s Early Warning for All agenda. This study highlights smart partnership into co-designing, co-developing, and co-implementing an impact-based EWS for geological risk in Jerai Geopark (Yan, Kedah), towards rejuvenating local resilience strategy through the development of DRR Yan Model and Resilience Living Lab in a national geological heritage area dominated by tourism industry. The key for successful community-led disaster risk reduction (CLDRR) lies in maintaining interest in resilience culture and motivation for local agenda at the grassroot level. We also demonstrate community-led DRR program with a unique localization strategy that addresses dam-related disaster risk. This study acknowledged that 40% of large dams in Malaysia are aging, necessitating new approaches to dam safety. Moreover, regional benchmarking for technological-based sociotechnical systems enabled by collaborative foresight and disaster informatics are a way forward to assess future emerging hazards, systemic risk, and compounding disaster. With good risk governance, evidence-based risk investment, and risk-informed decision making, as supported by all-of-society approach particularly in advancing a new partnership model for the public-private-academia-civil society, this study reports current demands for de-risk strategies that shall be systematically incorporated into decision-making, governance, and investments. The development of new strategies, actions, and initiatives are mutually explored towards inculcating targeted investments related to systemic risk reduction, and mainstreaming urban development planning of unattended risks should be made based on science, coupled by the Local, Traditional and Indigenous Knowledge (LTIK) approach. The de-risk investment efforts often jeopardize by a series of sudden, large-scale geological-induced disaster, resulting into the prolonged economic impacts continues to escalate and underscores the multi-scale investment for DRR agenda at a local level. By adopting a transdisciplinary approach to DRR and forward-looking risk-informed approach, this vulnerable region can further develop its resilience capacity to tackle complex challenges of climate risks. As a conclusion, risk-informed pathway in development planning, and a paradigm shift, can contribute towards promoting equitable and sustainable resilience in geologically risk sensitive regions.

Despite global initiatives, such as the Early Warnings for All initiative, operational realities lag behind when implementing people-centred Multi-Hazard Early Warning Systems (MHEWS) that consider multi-hazard interactions and take a first-mile, systems approach in the Global South.   This session will share perspectives from Practical Action, an international development organisation working in Latin America, Africa and Asia, to explore the diverse needs of the most at-risk and marginalised, and how core concepts of multi-hazard thinking integrate into different pillars and cross-cutting components of a MHEWS.  The session will highlight the mismatch between current ambitions and realities on the ground. Drawing on extensive experience from Practical Action, we will identify opportunities and challenges of moving towards MHEWS, emphasising the need for localised, inclusive strategies that genuinely address the diverse needs of the most vulnerable populations and fully encompass the meaning of multi-hazards, including hazard interrelationships, the dynamics of risk components, and the complexity of multi-hazard impacts. 

Background: The maternal mortality rate in Semarang is 121.5 per 100,000 live births, the second-highest in Central Java. The early warning system with the Early Warning Score and the maternal emergency early warning system (PDKM) still has various shortcomings to reduce MMR.Objective: This study aims to prove the effectiveness of the application of the PDKM Modified Early Obstetric Warning System (MEOWS) as an assessment of the risk of pregnancy in primary health facilities to reduce MMR in Semarang.Methods: The study was conducted on all pregnant women who came to Tlogosari Wetan, Tlogosari Kulon, Bandarharjo, and Bangetayu public health center in Semarang and were willing to participate in the study and were referred to government hospitals using national health assurance BPJS. Sampling was done by cluster random sampling by dividing the intervention and control groups. The study used a pretest-posttest control group design method by comparing the use of the MEOWS and the Poedji Rochjati Scorecard (KSPR) to the number of public health center referrals in Semarang. The data obtained will be analyzed statistically with the bivariate test, Mann-Whitney difference test, relative risk reduction, and absolute risk reduction.Results: The results showed that 21 of 43 (48.8%) patients were referred to the control group and 26 of 36 (72.2%) patients were referred to the intervention group. Mann-Whitney test of the number of referrals after the intervention within 3 months showed significant results (p = 0.033; p <0.05). There was an increase in the number of maternal referrals at the public health center in Semarang after the implementation of the MEOWS score by 1.48 times compared to using the KSPR (RR : 1.48 ; 95% CI : 1.02 – 2.13).Conclusion:The use of the MEOWS score can increase awareness of potential referrals and is associated with complications in patients.Background: The maternal mortality rate in Semarang is 121.5 per 100,000 live births, the second-highest in Central Java. The early warning system with the Early Warning Score and the maternal emergency early warning system (PDKM) still has various shortcomings to reduce MMR.Objective: This study aims to prove the effectiveness of the application of the PDKM Modified Early Obstetric Warning System (MEOWS) as an assessment of the risk of pregnancy in primary health facilities to reduce MMR in Semarang.Methods: The study was conducted on all pregnant women who came to Tlogosari Wetan, Tlogosari Kulon, Bandarharjo, and Bangetayu public health center in Semarang and were willing to participate in the study and were referred to government hospitals using national health assurance BPJS. Sampling was done by cluster random sampling by dividing the intervention and control groups. The study used a pretest-posttest control group design method by comparing the use of the MEOWS and the Poedji Rochjati Scorecard (KSPR) to the number of public health center referrals in Semarang. The data obtained will be analyzed statistically with the bivariate test, Mann-Whitney difference test, relative risk reduction, and absolute risk reduction.Results: The results showed that 21 of 43 (48.8%) patients were referred to the control group and 26 of 36 (72.2%) patients were referred to the intervention group. Mann-Whitney test of the number of referrals after the intervention within 3 months showed significant results (p = 0.033; p <0.05). There was an increase in the number of maternal referrals at the public health center in Semarang after the implementation of the MEOWS score by 1.48 times compared to using the KSPR (RR : 1.48 ; 95% CI : 1.02 – 2.13).Conclusion: The use of the MEOWS score can increase awareness of potential referrals and is associated with complications in patients.