1 © 2022 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of the World Bank with external contributions from experts at the Red Cross Red Crescent Climate Centre. The findings, interpretations and conclusions expressed in this work do not necessarily reflect the views of the World Bank, its Board of Executive Directors, or the governments they represent, nor of the Red Cross Red Crescent Movement. The World Bank does not guarantee the accuracy, completeness or currency of the data included in this work and does not assume responsibility for any errors, omissions or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes or conclusions set forth. The boundaries, colours, denominations and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of the World Bank, all of which are specifically reserved. Rights and permissions The material in this work is subject to copyright. Because the World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for non- commercial purposes if full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, the World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. 2 This report was prepared by the Red Cross Red Crescent Climate Centre (RCCC) and World Bank’s Global Crisis Risk Platform (GCRP). Authorship is shared by Tesse de Boer, Catalina Jaime, Juan Bazo, Cornelia Schulz, Rebeka Ryvola, Alex Sheaffer, Renate Meyer, Erin Coughlan de Perez, Evandro Holz, Mariano Rossi, Monica Dazzini Langdon, Facundo Palermo (RCCC) with inputs from Lindsey Jones, Ezequiel Miranda, Emma Philips and Indira Konjhodzic (World Bank). The team wishes to thank all organizations and individuals in Honduras who shared their experiences. Special thanks to the World Bank Honduras Country Team for feedback and inputs. Thanks also to Sarah Tempest for support with copy-editing. Funding for this research came the World Bank’s Crisis Risk Finance Analytics (CRFA) as well as the State and Peacebuilding Fund (SPF) in supporting the GCRP’s Multidimensional Crisis Risk Assessment and Monitoring Approaches program. The SPF is a global fund administered by the World Bank to finance critical development operations and analysis in situations of fragility, conflict, and violence. The SPF is kindly supported by: Australia, Denmark, France, Germany, The Netherlands, Norway, Sweden, Switzerland, The United Kingdom, as well as IBRD. Contacts: Tesse de Boer (boer@climatecentre.org) and Lindsey Jones (ljones12@worldbank.org). 3 Table of contents Abbreviations and acronyms ........................................................................................................ 5 Executive summary ...................................................................................................................... 6 Introduction .............................................................................................................................. 13 Methodology ............................................................................................................................. 14 Definitions and theoretical framing .................................................................................................. 15 Methods ............................................................................................................................................ 16 Data collection .............................................................................................................................. 16 Data analysis ................................................................................................................................. 17 Limitations..................................................................................................................................... 18 Findings ..................................................................................................................................... 19 Section 1: Disaster impacts ............................................................................................................... 19 Immediate impacts ....................................................................................................................... 19 Medium- to long-term impacts..................................................................................................... 21 Section 2: Drivers of risk ................................................................................................................... 24 2.1. Cascading and compounding hazards .................................................................................... 24 2.2. Dynamic pressures on vulnerability and exposure ................................................................ 26 2.3. Root causes of risk ................................................................................................................. 35 Section 3: Early warning signals, response and predictability of impacts ........................................ 41 3.1 Early warning and response for Eta and Iota ......................................................................... 41 3.2. Predictability of impacts ........................................................................................................ 47 Conclusions and recommendations going forward ...................................................................... 53 References................................................................................................................................. 56 Annexes..................................................................................................................................... 64 Annex 1: Overview of hurricanes Eta and Iota genesis, track, strength and direct impacts ............ 65 Annex 2: Hurricanes Eta and Iota forecast analysis .......................................................................... 69 Annex 3: Overview of disaster impacts Eta and Iota ........................................................................ 75 Annex 4: Pressure and Release Frameworks for main impacts of Eta/Iota ..................................... 78 Annex 5: Overview of available data sources for anticipation of crises in Honduras ...................... 80 Annex 6: Disaster risk reduction recommendations ........................................................................ 81 Annex 7: Data sources and additional outputs for mapping in this study ....................................... 82 4 Abbreviations and acronyms ACLED Armed Conflict Location and Event Data Project CENAOS Centro de Estudios Atmosféricos, Oceanográficos y Sísmicos (National Centre for Atmospheric, Oceanographic and Seismic Studies) CEPAL Comisión Económica para América Latina y el Caribe (Economic Commission for Latin America and the Caribbean) CODED/M/Ls Departmental, Municipal or Local Emergency Committees COPECO Comisión Permanente de Contingencias (Permanent Contingency Commission of Honduras) DRM Disaster Risk Management DRR Disaster Risk Reduction DTM Displacement Tracking Matrix ECMWF European Centre for Medium-Range Weather Forecasts ENSO El Niño–Southern Oscillation EWS Early Warning System FCV Fragility, conflict and violence FORIN Forensic Investigations of Disasters GBV Gender-based violence GCRP Global Crisis Risk Platform GDP Gross domestic product GloFAS Global Flood Awareness System HDX Humanitarian Data Exchange HRC Honduras Red Cross ICRC International Committee of the Red Cross IDB Inter-American Development Bank IFRC International Federation of Red Cross and Red Crescent Societies IP indigenous peoples KI key informant km2 square kilometres LGBTQI+ lesbian, gay, bisexual, trans, queer, questioning, intersex + LOCC (lack of) coping capacity MSF Médecins Sans Frontières (Doctors without Borders) NOAA National Oceanic and Atmospheric Administration NGO non-governmental organization NHC National Hurricane Center PAR pressure and release PAHO Pan American Health Organization RCCC Red Cross Red Crescent Climate Centre RQ research question SEPOL Sistema Estadístico Policial en Linea SINAGER Ley del Sistema Nacional de Gestión de Riesgos (Law on the National Disaster Risk Management System) SPS San Pedro Sula UNAH Universidad Nacional Autónoma de Honduras USD United States Dollar WASH water, santitation and hygiene WB World Bank WFP World Food Programme 5 Executive summary The dual Tropical Storms Eta and Iota that hit Honduras in November 2020, are an example of how natural, socioeconomic and political drivers can produce compounding impacts. Compound crises such as these hinder, and sometimes even reverse, development gains as direct losses and damages and long-term implications for fragility and macroeconomics affect poverty and prosperity. This retrospective analysis explores these risk interactions along with the available warnings and preventive actions, to draw lessons for future crises of a similar nature as part of the Global Crisis Risk Platform (GCRP) of the World Bank. This retrospective case study draws on compound risk analysis methodologies and multi-hazard early warning system designs to understand what the underlying drivers of risks were at the time, and how these could be anticipated in the future. The case study draws on peer reviewed literature and key informant interviews, publicly accessible data and geospatial analysis to consider compounding and cascading risk interactions in 2018–2020 in Honduras, their attendant impacts and risk drivers, and available warnings as well as the communication and early actions associated with them. Based on the findings of the compound risk and early warning capacity analysis, this report recommends the following actions: • Prioritize the development of an integrated risk monitoring system, that integrates available dynamic and static information on exposure and vulnerability. Preceding hazards (drought, dengue, COVID-19, floods) and underlying deep vulnerabilities related to the economic situation, violence and migration in Honduras all compounded the impacts of Eta and Iota. In the future, a continuous monitoring of dynamic pressures, in addition to the monitoring of natural hazards along with the root causes of pre-existing vulnerability, is crucial to understand who is most at risk when hazards of the intensity of Eta and Iota face Honduras again. Despite the Disaster Risks Reduction efforts of the Government of Honduras, during November 2020, the lack of a multi-hazard early warning system approach severely limited the ability of actors in the country to foresee the humanitarian impacts. The hazard interactions were not properly recorded in an integrated, standardized and coherent manner, and social vulnerability information was not part of the DRM system. When Tropical Storm Eta hit first, there was anecdotal understanding of which areas were more at risk, yet there was not a fully organized system to support COPECO, different ministries, local communities, humanitarian and development actors to take more effective decisions in advance of the shocks. A framework that integrates information on food security, health, violence, housing and migration (in addition to hazard forecasting and monitoring) could have been instrumental in anticipating the impacts during and after Eta and Iota in 2020. The report details suggested indicators to include, based on the identified compounding dynamics. • Improve modalities to monitor violence as an essential component of disaster risk management and integrated (dynamic) risk monitoring. Violence should be considered an integral part of risk monitoring and decision-making for future crises. Urban and rural violence related to maras and drug trafficking groups have been shown to deeply affect all dimensions of risk in the population including exacerbating vulnerabilities, contributing to migration, altering cognitive risks bias, and increasing risky behaviours such as resistance to evacuation. The geospatial analysis in this research shows that the areas most impacted by Eta and Iota also suffered from some of the highest homicide rates. People in areas affected by violence 6 also have less access to Government and other humanitarian and development support, predisposing them to a vicious cycle of disaster impacts. Monitoring violence can improve decision-making in times of crisis. A dynamic risk system would inform if, and to what extent, the levels of violence have repercussions for the capacity to act before a shock and identify subsets of the population that need special protection considerations in times of evacuation. Agencies such as the International Committee of the Red Cross could play a key role on this process. • Expand research on compound risks to guide policy- and decision-making. The evolution of risks over time and across geographies is complex, and impacts from crises such as Eta and Iota, in combination with COVID 19 impacts, and the several preceding disasters induced by drought, floods and dengue, can only be fully understood by analysing the accumulation of risk over time and space. A greater investment in continued research, practice and policy (ideally at sub-national scale) on compound risks including scenarios under a changing climate, will enable the Government of Honduras and its partners to fully understand the dynamic changes in risks, with emphasis on supporting marginalized and poor populations, migrants and those living with violence. • Data quality, granularity and accessibility for hazards, exposure and vulnerability indicators needs to be improved and integrated. Data improvement is one the cornerstone of an open, safe, effective and harmonized system that can reduce long term risks and anticipate a crisis of similar magnitude in the future. Findings suggest that available information was also not used in the preparedness and response coordination at the time, signalling the need for the institutionalization of the use available science and risks data for decision making. • Root causes and structural faults need to be addressed to reduce the risks of disasters. This analysis shows the importance of considering compound risk, especially hazard exposure, violence and vulnerability together. For any work on decision-support tools, it remains crucial to holistically address both the drivers of risk as well as the root causes. The unsafe conditions and root causes leading to high levels of vulnerability produced the compounding crisis observed during and after Eta and Iota – with high humanitarian needs as a result. While more effective early warning early action and disaster response can reduce these impacts in the future, long-term disaster risk reduction and climate change adaptation measures are the key to address the root causes of risks, including environmental degradation, the planning of human settlement and agriculture, the limited public investment in the maintenance of critical infrastructure and systems, and the systems to support transparency and accountability at all levels. • Establish the next generation of impact-based forecasting systems. A major gap in the early warnings of Eta and Iota, as well as for previous events, was the lack of public guidance on the potential disaster impacts and suggested early actions for imminent hazards. Warnings for Eta focused on windspeed and hurricane track and overlooked the high risk of flooding and landslides in the north-west of the country. The heavy rainfall was forecast well in advance, even before the hurricane forecasts, but this window of opportunity was not utilized, resulting in late evacuations, poor preparedness and avoidable loss of lives and livelihoods. A comprehensive impact-based forecasting system would help to facilitate warnings and advance support to the most at-risk communities, thereby minimizing the disaster impacts. Such a system would build on an in-depth understanding of hazard interaction and underlying vulnerability to support decision-makers and the public in taking actions based on likely impacts arising from forecast hazards such as a hurricane. COPECO, including CENAOS, should be supported to shift towards the use of impact-based warnings. Warning systems should be coupled with campaigns to increase the public’s trust in government entities. Civil society and 7 humanitarian actors such as the Honduras Red Cross can play a role in growing trust in public institutions through their strong community networks. • Enhance coordination protocols and support local capacity building for disaster response, including the financial capacity to act when required. Lastly, in the Honduran system where municipalities are the executive arm of the disaster management system, reforms are required for improved communication, better trained and motivated people in key monitoring positions, and funding to maintain these structures and act when needed. Additionally, a platform of coordination between disaster risk management, humanitarian, development, climate and peacebuilding/violence actors is key for climate change adaptation and disaster risks reduction plans, including anticipatory action. Such reforms should also ensure that warnings to key players and the public are timely, convey impacts and provide guidance on preventative actions. In addition to long term risks reduction investments, warning communication and dissemination protocols are crucial to reduce residual risks. A summary of key findings from the retrospective case study is presented below. Affecting the same geographical area in close succession, Tropical Storms Eta and Iota resulted in cascading hazards culminating in widespread flooding and landslides, especially affecting the north and north-west of Honduras (Figure ES2). As a result, communities living close to the major rivers in the north, especially in the Sula Valley, saw their neighbourhoods flooded – resulting in drowning, damage and destruction of infrastructure and cutting off many communities by damaging roads and communication networks. Landslides on the hills at the periphery of San Pedro Sula and in the west of Honduras also resulted in widespread damage. Economic losses and damages are estimated at 1.9 billion United States dollars, with a 0.8 per cent reduction in economic growth in Honduras during 2020 (World Bank, 2021a), deepening the damage provoked by the COVID-19 pandemic. Almost 80 per cent of economic damages were accounted for by the productive (44 per cent) and social (35 per 8 cent) sectors. The most affected productive subsectors were commerce and industry, with 34 per cent of the total damages, and agriculture with 8 per cent. In the case of the social sector, housing accounted for 26 per cent of the total damages. Infrastructure damage accounted for 15 per cent of the total, with the most affected subsector being transportation (9 per cent of the total damages). These results show that families' livelihoods as well as the locations where they lived were severely affected (CEPAL, 2021; p. 21). Shelters could hardly absorb the large, displaced population after Eta, and the system was overwhelmed when Iota extended the floods further. Overcrowding and loss of essential infrastructure for water, sanitation and health services resulted in rising COVID-19 cases, dengue and gastro-intestinal diseases. Loss of employment in the affected urban areas, along with loss of (subsistence) crops and labour options in rural areas, contributed to rising food insecurity by limiting households’ ability to buy food. The damage to infrastructure and agricultural lands from the floods and landslides resulted in major economic losses – compounding the pre-existing economic recession in Honduras as a result of COVID-19. • Areas affected by violence and the economic consequences of COVID-19 were hardest hit by Eta and Iota, partly due to physical conditions but mainly because of the dynamic pressures on vulnerability and exposure that accumulated over time (Figure ES1 and Figure ES3). The unsafe conditions that directly produced the impacts observed during and after Eta and Iota were a product of underlying dynamic pressures, influenced by preceding crises and structural issues (Figure ES3). Data gaps persist. In a context where monitoring and evaluation primarily focus on urban departments such as Cortés, Atlántida and Sánta Barbara, these coastal and predominantly rural communities (e.g., La Mosquitia) remain understudied and left behind. o The dual Tropical Storms compounded the existing economic crisis as a result of COVID-19, political instability and corruption that had already severely increased 9 economic and socio-political vulnerability. Trust in Government was low, financial options for risk reduction were limited and poverty and inequality had risen sharply in the period before the Tropical Storms hit. In response to the COVID-19 lockdown, violence dynamics had shifted in these areas – with gangs intensifying their extortion activities in the latter part of 2020. Furthermore, especially in the urbanized north – most affected by the floods and landslides during Eta and Iota – the export industry, agriculture and service economies were hit hard, resulting in widespread loss of income especially in the informal sector. o Dynamic migration processes in Honduras result in more people in at-risk areas – especially peri-urban areas – negatively affecting disaster preparedness and action. ▪ Violence, displacement and urban exposure: violence permeates through all aspects of daily life in urban and rural areas in Honduras, particularly affecting physical safety, economic vulnerability (through extortion) and eroding socio- political capacity in the country. In both rural and urban contexts, displacement frequently occurs in response to the threat of recruitment or extortion by gangs. While displacement is often a life-saving strategy, these individuals tend to end up in more insecure areas. In the case of San Pedro Sula, the neighbourhoods severely affected by the storms were some of the most insecure and poorest areas – where access for humanitarian actors is limited. ▪ Drought, food insecurity and migration: Honduras suffered a severe multi- year drought in the Dry Corridor between 2014–2019. Limited livelihood diversification and a high dependence on agricultural daily labour for the poorest or subsistence farming for the slightly better off, meant that the drought severely eroded rural livelihood opportunities and increased food insecurity. Within the Dry Corridor, the prolonged drought served as a push factor for rural-urban and international migration. However, rapid urbanization into areas ill-equipped to receive population influxes created the conditions for expanding informal settlements onto floodplains and unstable slopes, increasing exposure to floods and landslides, and limiting income- generating opportunities especially in the informal sectors. o Secondary impacts relating to health (dengue, COVID-19, gastro-intestinal diseases, limited access to care for those already in need, such as pregnant women) were not only caused by the damage and disruption from Eta and Iota, but occurred in a context that was still recovering from a severe dengue epidemic (2019) and the ongoing COVID-19 pandemic. These peaks in demands exerted pressure on an already strained health system suffering from structural underfunding, corruption and inequal distribution of services. • Root causes of vulnerability (Figure ES3) predispose Honduras to similar crises in the future. The combined result of unsafe buildings, settlement in at-risk areas, poverty, inequality and ineffective governance increases dynamic pressures on unsafe conditions – including a lack of investment in, and maintenance of, disaster risk reduction interventions as well as health and social security systems that may help communities to cope with such hazards. Furthermore, ongoing environmental degradation, especially deforestation, soil erosion and the accumulation of waste – partly because of climatic factors, but also due to Government and economic processes – have been shown to intensify flooding and landslide onset and aggravate impacts. These dynamics affect groups of the population differently, largely determined by social and economic exclusion, inequality and marginalization. 10 • In addition to the root causes of vulnerability, unsafe conditions were aggravated by a lack of early action, mainly because of a lack of understanding of potential impacts, limited dissemination of warnings and lack of protocols, plans and finances to execute anticipatory action. Especially limited and late evacuation resulted in preventable loss of lives and livelihoods. People were either not reached with warnings and evacuation support, or were unwilling to leave. The floods affected some of the most violent areas of Honduras, and the fear of being robbed made people unwilling to leave their houses in advance of the disaster. Furthermore, sexual abuse and gender-based violence rates were high in shelters and the fear of violence prevented people from evacuating to shelters. A limited understanding of the impact levels along with gaps in the coordination and guidance for evacuation added to this challenge. Furthermore, in the weekend before Eta stormed over Honduras there was the long-awaited national holiday Semana Morazánica, which had been postponed due to the COVID-19 lockdown. The holiday was a major driver of the overall low preparedness of Honduras for Eta – with many people on the move to see family and authorities out of office or occupied with the logistical organization of the holiday. Nonetheless, there were examples of effective local responses to Eta and preparedness for Iota, with civil society organizations maintaining communications with affected communities, mobilizing resources and distributing emergency support. The lack of preventative action based on forecasts ahead of the crisis was also attributed to weak coordination between various levels of government, which comes back to the limited capacity at municipal level. • At the time of the crisis, there was no multi-hazard and impact-based early warning system in place, and this severely limited the ability of actors in the country to foresee the humanitarian crisis in November 2020 and beyond. A major gap in the early warning for Eta and Iota, as well as for previous events, was the lack of public guidance on the potential disaster impacts and suggested early actions for imminent hazards. Warnings for Eta focused 11 on windspeed and storm track and overlooked the high risk of flooding and landslides in the north-west of the country. The heavy rainfall was forecast well in advance, even before the hurricane forecasts, but this window of opportunity was not utilized, resulting in late evacuations, poor preparedness and avoidable loss of lives and livelihoods. While a full impact-based system takes time to develop, the limited communication on flood risk is a missed opportunity, given the available forecasts and understanding of the basins most at risk. • Overall information on the drivers of risk – including preceding disaster events, violence, migration and socioeconomic influences – were not considered ahead of Eta and Iota. Information sources were not utilized to their full potential and monitoring systems were incomplete, resulting in crucial knowledge gaps at the time of the crisis in November 2020. For example, Honduras faced a severe humanitarian crisis due to drought, induced by ENSO, between 2014 and 2018. This crisis heightened food insecurity levels, inducing malnutrition, yet those facing the direct and indirect impacts of the drought were not tracked and supported to cope with the subsequent hazards. This accumulation of risks was compounded by the disruption to the health system due to the dengue epidemic in 2018/9, followed by the agriculture damages due to floods in 2018 and early 2020, and the extreme impacts across all sectors of COVID-19. Understanding population movement within Honduras at this time along with the key vulnerabilities and lack of coping capacities of rural families would have helped to better understand the impacts of the tropical storms. The systemic root causes of vulnerability and exposure are likely to worsen without adequate interventions, especially under a changing climate. • Potentialize capacities of the Honduras population and local actors. Despite all the unsafe conditions and complexities faced by Hondurans and international migrants, it is crucial to highlight the extraordinary capacity that the population has to recover from crisis. Investments in long term disaster risks reduction and anticipatory action has the opportunity to build on existing local and municipal practices by community based organizations, Non- Governmental Organizations and institutions such as the Honduras Red Cross, which has branches in the most at risks areas of Honduras. Tropical storms Eta and Iota stand out as a crucial moment in Honduras’ history. While Honduras is prone to extreme weather events including droughts, floods, hurricanes and landslides, the tropical storms were foregrounded by longstanding socio-political, economic and development challenges including: contentions over the 2017 presidential elections and subsequent corruption scandals; a multi-year period of drought; the largest dengue fever outbreak; the spread of COVID-19; and some of the world’s highest rates of violence in 2020 (Figure ES1). In this fragile context, Eta and Iota impacted more than 4.5 million people. It is important to invest in long term disaster risk reduction, including approaches to anticipate the impacts of compounding events like Eta and Iota along with root causes and drivers of risk to reduce risks and be prepared for the future. Global climate change projections indicate that hurricanes will become stronger, move slower and bring more heavy rainfall to Central America. This means that hydrometeorological events with the characteristics of Eta and Iota are very likely to become more common in the future. The impacts of the compound crisis brought about by Eta and Iota, in combination with COVID 19 can only be fully understood by also considering existing vulnerability and the accumulation of risk. The findings and recommendations of retrospective analyses in contexts experiencing fragility, conflict and violence, can form a critical basis for risk-informed climate adaptation, risks reduction, early action, response and recovery. 12 Introduction During a compounding crisis, multiple hazards over space and time interact to produce impacts that are larger than the sum of its parts (de Ruiter et al., 2020). There is emerging recognition of the importance of understanding compounding impacts. Traditional single-hazard approaches fail to capture and predict these complex dynamics (Cavallo & Ireland, 2014; de Ruiter et al., 2020; Marchezini & Wisner, 2017; Oliver-Smith et al., 2016; UNDRR, 2022; Zscheischler et al., 2020). One such example is the complex humanitarian crisis in Honduras during and after November 2020, when Tropical Storms1 Eta and Iota ravaged the country in a span of two weeks. Despite general awareness of potentially compounding impacts in situation reports (CEPAL, 2021; IFRC, 2021a), there remain knowledge gaps on exactly how, when and why the various risks in Honduras combined to produce the impacts observed during and after Eta and Iota. This retrospective analysis explores these risk interactions along with the available warnings and preventive actions, to draw lessons for future crises of a similar nature under a changing climate. The Republic of Honduras is a country situated in Central America. It is bordered by Guatemala, El Salvador and Nicaragua, as well as water borders on the Pacific Ocean and the Caribbean Sea. Its largest city is the capital Tegucigalpa. The country spans about 112,090 square kilometres (km²) and has a population of 10,062,994 inhabitants (CIA, 2022; World Bank, 2022b). The main economic sectors in urban areas are industry and services, especially around the capital Tegucigalpa and the second largest city San Pedro Sula. Remittances form an important source of income nationally (World Bank, 2022c). Agriculture and related agri-businesses remain the main livelihood source for rural and poor populations. Over 65 per cent of Hondurans live in poverty and over 70 per cent of the population works in the informal sector - which contributes to 15–25 per cent of gross domestic product (GDP) (World Bank, 2022c). Exports are also important for the economy – the country is rich in natural resources including coffee, minerals, sugar cane and tropical fruit. Politically, the country benefited from high political stability until the coup d’état in 2009. However, controversies surrounding the 2017 presidential election and several major corruption scandals, along with some of the world’s highest rates of violence have eroded the political stability in Honduras (Human Rights Watch, 2020). Honduras is frequently affected by hurricanes, tropical storms, drought, floods and landslides, although major disasters occur less frequently (USAID, 2017). In 1974, Hurricane Fifi resulted in 5,000– 8,000 deaths; and, in 1998, over 5,000 people lost their lives and over 8,000 went missing due to Hurricane Mitch (Zambrano et al., 2021). Tropical Storms Eta and Iota passed over Honduras and neighbouring countries in November 2020 in quick succession. While mortality during and after the “dual tropical storms”2 was far less (96 people), compared to Fifi and Mitch, the number of people affected is estimated at 4.5 million in Honduras – almost half the population at the time (IFRC, 2020b). The Tropical Storms took place during a critical time, when Honduras was grappling with the COVID- 19 pandemic and its economic fallout, recovering from prolonged drought, and still fighting the worst 1 Eta and Iota made landfall in Nicaragua as Category 4 hurricanes. When they hit Honduras, Eta was a tropical depression and Iota was a tropical storm. After crossing Honduras, both continued as tropical depressions. We will refer to these two events as “Tropical Storms” in the report. It should be noted that the hurricanes' precipitation bands already had an influence on rainfall accumulations before the Eta and Iota made landfall, as will be explored further in Annex 1. 2 Eta and Iota passed over Honduras in the span of two weeks. As the storms did not develop at the same time (which would be classed as “twin hurricanes”), we will use the term “dual tropical storms” as they did affect the same geographical area in quick succession. 13 dengue outbreak in decades. Honduras is affected by intense violence including street gang-related violence, drug trafficking, international organized crime, gender-based violence (GBV), state violence and political repression (IDMC, 2016). The combination resulted in a complex humanitarian crisis (IFRC, 2020b). Despite general awareness of potentially compounding impacts, there remain knowledge gaps on exactly how, when and why the various risks in Honduras combined to produce the impacts observed during and after Eta and Iota. Thus far, there has not been an in-depth analysis to reconstruct the accumulation of risk, identify oversight in the response and highlight opportunities for improvement. The World Bank initiated a retrospective case study on the humanitarian crises around Tropical Storms Eta and Iota as part of the Global Crisis Risk Platform (GCRP). The World Bank GCRP aims to anticipate, track and respond to risks related to complex crises (World Bank Group, 2018). A core aspect of GCRP is the development of multidimensional crisis risk analytics, with a workstream centred on three core components. These include the deployment of i) a suite of multidimensional crisis risk analytics (including a Compound Risk Monitor); ii) regular multidimensional crisis risk horizon scans; and iii) a series of retrospective compound risk country case-studies. Work outlined in this project relates to the latter activity, seeking to analyse and draw insight from past compound risk events to help inform risk analytics in the context of the GCRP. In this report, the Red Cross Red Crescent Climate Centre (RCCC), with the support from the World Bank, explores the various compounding events that occurred in Honduras before and during the recent Tropical Storms and the resulting complex humanitarian disaster. In the two years before, several interconnected hazards occurred that affected the populations’ exposure, vulnerability and lack of coping capacity as well as the effectiveness of the early warning system. This retrospective analysis of Tropical Storms Eta and Iota in Honduras aims to address the following research questions: 1. What were the underlying drivers of risk that led to the compound crisis (including relevant natural, socioeconomic and political factors), and how did the various drivers interact over time? 2. What early warning signs were apparent ahead of the compound crisis, and to what extent were they acted upon by relevant authorities? 3. To what extent was the compounding nature of the crisis predictable, and what lessons can be learned for future? 4. What decision support tools would have been needed to support effective anticipatory actions ahead of the compound crisis (including relevant political windows of opportunity)? The compound risk case study draws on qualitative primary data gathered through key informant (KI) interviews, a geospatial data analysis, forecast analysis and a literature review. The findings discuss the key impacts seen during the crisis, performance of the early warning system, and the drivers and root causes of risk. These serve as the building blocks for the assessment of predictability of the main impacts as well as the needs for future decision support. Methodology This retrospective disaster analysis takes a case study approach, drawing on both primary qualitative data from KI interviews and quantitative and qualitative data extracted from secondary sources, to analyse the evolution of risk that culminated in the humanitarian crises during November 2020 in Honduras. It focuses on events before the crisis to understand how, why and where compounding 14 dynamics produced the impacts observed. The study aims to understand the interactions between various hazards, multi-dimensional vulnerabilities and impacts observed during Tropical Storms Eta and Iota and to understand how and to what extent the population was affected by aggravated disaster impacts due to multiple hazards and violence. It also analyses the early warning system and disaster management response to assist future investments in decision support in Honduras. Definitions and theoretical framing Understanding risk is essential to accurately anticipate crises. Risk factors (hazards, vulnerability and exposure), risks drivers (e.g., violence, rapid urbanization, environmental degradation and climate change) and root causes (e.g., corruption and inequality) influence the outcome of a crisis (Wisner et al., 2014). Without understanding these complex interactions across time and geographies, it is not possible to develop policies, strategies and measures that can reduce risks, manage residual risks, and respond and recover to disasters effectively. Compound risk analysis offers an opportunity to understand the potential impacts that disasters can have in society (UNDRR, 2022). The methodology was influenced by research on cascading and compounding hazards and impacts (Cavallo & Ireland, 2014; de Ruiter et al., 2020; Pescaroli & Alexander, 2016; UNDRR, 2022). The in-depth assessment of early warning signals for the Tropical Storms and associated crises draws on the Forecast-based Financing manual methodology (GRC et al., 2020) and is structured along key elements of the World Meteorological Organization’s Early Warning System: understanding risks, forecast monitoring, warning dissemination, and response (WMO, 2018). This study uses the following definitions: Hazard A process, phenomenon or human activity that may cause loss of life, injury or other health impacts, property damage, social and economic disruption or environmental degradation (UNDRR, 2022) Vulnerability The conditions determined by physical, social, economic and environmental factors or processes which increase the susceptibility of an individual, a community, assets or systems to the impacts of hazards (UNDRR, 2022) Exposure The situation of people, infrastructure, housing, production capacities and other tangible human assets located in hazard-prone areas (UNDRR, 2022) Compound events Events that either occur simultaneously or successively combining into severe impacts; extreme(s) combined with systemic conditions that amplify impact; or extremes that result from combinations of ‘average’ events (Pescaroli & Alexander, 2016; UNDRR, 2022) Cascading events Extreme events where effects increase in progression over time due to cascading events, with strong impact and unexpected secondary events as a result (Pescaroli & Alexander, 2016) Risk Drivers Processes or conditions, often development related, that influence the level of disaster risk by increasing levels of exposure and vulnerability or reducing capacity (UNDRR, 2022) Unsafe conditions Immediate causes of vulnerability (Oliver-Smith et al., 2016) Dynamic Immediate processes that translate root causes into unsafe condition, typically pressures/processes changes on the order of magnitude of a decade or two or three but influenced by risk drivers (Marchezini & Wisner, 2017; Wisner et al., 2014) Root causes of The root causes of vulnerability involve social and economic structures, such as the vulnerability characteristics of power, wealth and resources distribution, as well as ideologies and historical heritage. Such root causes may change, albeit slowly (Marchezini & Wisner, 2017) 15 Methods Given the scope of the research questions, time and data access limitations, the study uses qualitative methods, including content analysis, complex risks interactions mapping and geospatial analysis (see Figure 1). The research follows an iterative analysis approach, of which the steps are outlined below for data collection and analysis. Figure 1: Overview of data collection and analysis process. Data collection First, inspired by the Forensic Investigations of Disasters (FORIN) framework (Oliver-Smith et al., 2016), a comprehensive literature review was carried out on the following topics: • Disaster impacts of Tropical Storms Eta and Iota across different sectors based on national reports, humanitarian situation reports, appeals and evaluations. • Scoping of disaster impacts of natural hazard-induced disasters in 2018–2020, based on ReliefWeb and EM-DAT records of previous disasters; followed by identification of their respective impacts, early warning early action, and risk drivers based on secondary literature. This was summarized in the disaster timeline. Only disasters that included information on negatively impacted populations were included in the disaster timeline. The results were visualized using Miro and included as a timeline graph in this study. • Drivers and impacts of the situation of violence as well as migration dynamics across Honduras. • Forecast information and warning communication for Eta and Iota, including news reports in English and Spanish and official government communications. Furthermore, based on the identified natural hazard-induced disaster events (7 events, sources summarized in Annex 7), geospatial data was analysed to identify key datasets that could contribute to the risks analysis. Open-source geospatial datasets were used to generate the maps. Datasets were researched through an online scoping review of open-source data portals (e.g., HDX) as well as Google 16 search on specific topics. A complete list of researched datasets on Honduras can be found in Annex 7. The thematic maps were created with ArcMap Desktop (Version 10.8) including cleaning data, visualizing relevant information and exporting maps through the layout function. Data included: areas affected by disaster events in 2018–2020; violent incidents 2018–2020 (ACLED); Sistema Estadístico Policial (SEPOL) monthly homicide data (SEPOL, 2022); and municipal-level data on vulnerabilities (INFORM Index) along with lack of coping capacity (INFORM Index) to support the risks analysis process3. Based on the preliminary understanding of Tropical Storm impacts, EWS and general context in Honduras primary qualitative data was collected through KI interviews, focusing on the four research questions, and specifically targeting identified gaps in the literature. Interviewees were selected based on their knowledge of the disaster events (impacts and risks factors), understanding of the violence and migration context, and experience in early warning and early action in Honduras. Selected key informants represented a range of international humanitarian organizations including the UN, Honduran civil society organizations, academia, government actors, and the World Bank. A total of 23 KIs were interviewed in Spanish via an online platform. Interviews were transcribed and coded by two separate teams based on the key themes for each research question. Interviews were anonymized and affiliations generalized to “humanitarian worker”, “civil society organization” (including the Honduras Red Cross), “government”, “development actor” and “academic”. Additionally, three focus group discussions were organized with experts in the field of Disaster Risk Management in Honduras, while anticipatory action nationally and globally helped to validate interview data and identify opportunities for improvement in the future. Data analysis The data analysis for the four research questions included an analysis of the Tropical Storms’ impacts and drivers of risk as well as a review of the early warnings systems at the time. Content analysis as well as the coding of interviews formed the basis of the analysis: • Synthesis of the main disaster impacts in short- and medium- to long-term. • Reconstruction of hazard interaction to identify cascading and compounding dynamics, including analysis of the groups, geographical areas and sectors most affected to understand unsafe conditions resulting in the impacts. • Analysis of major social, political and natural-hazard related events in 2018–2020 and their influence on vulnerability (socio-political, economic, environmental, physical) and exposure to the dual Tropical Storms to understand dynamic pressures and accumulation of risk, visualized in a disaster timeline. A hotspot map was created to identify geographic areas where impacts of multiple hazards, violence, high vulnerability and lack of coping capacity coincided geographically. We then used population density to understand the areas with highest exposure. • Identification of dynamic pressures and root causes of unsafe conditions, through development of several pressure and release (PAR) analyses for the major impacts – 3 INFORM Index developed by the Joint Research Centre of the European Commission (Marin-Ferrer et al., 2017). For this study, the INFORM index was used to analyse levels of vulnerability and lack of coping capacities of the municipalities by integrating it in the geospatial analysis. The index provides a score of 1–10 for vulnerability and lack of coping capacity (0 – very low; 10 – very high). A list of the data and sources used as indicators in this analysis can be found in Annex 7. 17 summarized in a general PAR Figure 12. This analysis draws on the content analysis and expert discussions within the research team with experts. To understand the effectiveness of early warnings at the time as well as opportunities for future anticipatory action, we conducted a forecast analysis and analysed the warning communication and dissemination process at the time. The forecast analysis follows the National Hurricane Center (NHC– NOAA) methodology (National Hurricane Center Center & Central Pacific Hurricane Center, 2020) and draws on open-source databases of the forecasts at the time. This analysis offers detailed forecast information for Eta and Iota and includes a description of existing early warnings for the other hazards in the study. The analysis of early warning dissemination, communication and response for all the hazards in study, consisted of coding and content analysis of all the literature and interviews. Expert discussions among the study team took place to identify salient points and the deliberation of key recommendations. The analysis was inspired by the Forecast-based Financing approach (GRC et al., 2020); specifically, we analysed the potential for Impact-based Forecasting (Harrowsmith et al., 2020). This study proposes options for anticipating the impact of similar crises in the future (RQ 3), based on the reconstruction of the accumulation of risk over space and time in Honduras in the years preceding the 2020 crisis (RQ 1) and our understanding of the early warning capacity at the time (RQ 2). Suggested options to address observed gaps were compiled based on a scoping of available data and ongoing initiatives in Honduras. The proposed set-up of such a system draws on the Impact-based Financing approach (GRC et al., 2020; WMO, 2018). Recommendations for decision support (RQ 4) build on the findings for the other three questions, exploring how drivers of risk and root causes can be addressed – along with needs for operational disaster response and recovery. Content analysis and a focus-group discussion were used to synthesize findings and recommendations. Limitations This research has several limitations that should be considered when interpreting the results. While the experts interviewed in this study had lived through the events around Eta and Iota, it should be noted that there may be some recall biases. This retrospective study is carried out two years after the crisis, and therefore was limited by what people could remember. Furthermore, the content analysis draws on news reports as well as academic and grey literature, which in Honduras has a strong focus on urban areas. It was a challenge to access KIs and secondary data to corroborate narratives from rural and remote areas of impacts and experiences in the years before the Tropical Storms. Furthermore, there are some important gaps in the mapping exercise that are considered critical for the overall analysis. Firstly, there is a lack of (geospatial) data for several key vulnerable groups: displaced people, both due to the Tropical Storms and other events and conflict, and the locations of remote, indigenous communities in sparsely populated departments. This would be valuable information to add. Secondly, the analysis is partly based on OpenStreetMap and especially the east of Honduras seems poorly mapped so far, thereby potentially missing opportunities to better understand hazard exposure in these areas. More granular geospatial information on the various flooding events (including Eta and Iota), especially in the Cortés Department, and information on violence on a lower scale than municipalities could support analyses of the key neighbourhoods affected. 18 Findings The findings first explore the impacts of Tropical Storms Eta and Iota (RQ 1) in Sections 1 and 2. Section 1 outlines the disaster impacts and unsafe conditions observed during and after Eta and Iota (November 2020). In Section 2, hazard interaction, drivers of vulnerability and root causes are explored. Lastly, Section 3 analyses the performance of the early warning system and the humanitarian response, and opportunities for the future anticipation of impacts (RQs 2 and 3). Research question 4 is answered through the recommendations. Tropical Storm Eta entered Honduras on 4 November 2020, bringing extremely intense winds and heavy rainfall. The rainfall concentrated in the north(west) of Honduras, resulting in extensive riverine and flash flooding in several valleys (most notably the Sula Valley) along with landslides across the north. Iota entered Honduras on 17 November and heavy rains affected the same geographical areas, thereby intensifying the floods that had not yet receded from Eta. A technical summary of the Tropical Storms’ track and strength as well as an overview of flood extent after Eta and Iota can be found in Annex 1. Section 1: Disaster impacts According to an evaluation by the Economic Commission for Latin America (CEPAL) in 2021, the combined effects of both Tropical Storms directly impacted 437,000 people4 and indirectly affected 4.5 million people, including the (temporary) displacement and evacuation of 937,000 persons (CEPAL, 2021; Comisión Permanente de Contingencias (COPECO) as sourced by IFRC (2021a), IDMC, 2020). The crisis left 2.8 million people in need of humanitarian support (IFRC, 2021a). Economic losses and damages are estimated at 1.9 billion United States Dollars (USD), with a 0.8 per cent reduction in economic growth in Honduras during 2020 (World Bank, 2021b), deepening the damage provoked by the COVID-19 pandemic. Almost 80 per cent of economic damages were accounted for by the productive (44 per cent) and the social (35 per cent) sectors. The most affected productive subsectors were commerce and industry, with 34 per cent of total damages, and agriculture with 8 per cent. In the case of the social sector, housing accounted for 26 per cent of total damages. Infrastructure damage accounted for 15 per cent of the total, with the most affected subsector being transportation (9 per cent of the total damage). These results show that families' livelihoods and the place where they live were severely affected (CEPAL, 2021, p. 21). Below we discuss several of the main impacts observed. A full summary of the impacts can be found in Annex 3. Immediate impacts Death by drowning or burying. 95 people lost their lives in addition to ten missing persons as a direct result of the Tropical Storms, mainly from drowning in the floods during the Tropical Storms or during building collapses because of landslides in Cortés, Santa Bárbara and Lempira districts (CEPAL, 2021). 4 Please note, these impact measurements vary slightly across agencies (COPECO and CEPAL) due to limitations in baseline information (generated from the latest population consensus in 2013), thus leaving necessary reliance on alternative sources generated by various institutions. Additionally, there are different working definitions and concepts of “affected persons” that impact the statistics (CEPAL, 2021). The figure on displacement includes those evacuated and sheltered following COPECO, although the methodology could not be verified and data is not disaggregated by gender, age and type of displacement. 19 Of the ten missing people, two were fishers (CEPAL, 2021). The most fatalities occurred in Cortés (32), Santa Bárbara (16) and Lempira (12). Isolation of communities. Communications were cut off to more than 95,000 people in just Copán during and after Eta, which increased after Iota (OCHA, 2020a)5. A total of 927 roads were affected and more than 72 bridges damaged and 62 were fully destroyed. As roads and bridges were washed away, communities (over 300,000 people) were isolated for several weeks (IFRC, 2021a). Loss and damage of homes resulting in displacement under precarious conditions . Statistics on evacuation differ, but the CEPAL (2021) states that 170,000 people were evacuated. As a result of floods and landslides, over 90,000 homes were damaged (CEPAL, 2021). The rivers with the largest floods were in the Sula Valley (rivers Ulúa and Chamelecón), Choluteca River in the south, Aguán Valley in Colón and the Leán Valley in Atlántida (IFRC, 2021a). Flood extent maps in Annex 1 show the areas affected in more detail. OCHA (2020) statistics indicate that 88,000 people needed temporary shelter and 35,000 people needed house repairs. CLAC reports that more than 56.000 people were sheltered in official locations although this number does not reflect people staying with friends and family ((CLAC, 2020). In shelters and affected areas in general, several health and protection issues emerged. The spread of COVID-19 increased with San Pedro Sula as one of the highest reported cases, especially because of the overcrowded and undersupplied shelters and lack of personal protection equipment (IDB, 2021; UNICEF, 2020). In the shelters, women and girls were exposed to sexual abuse, and gender- based violence (GBV) rates increased markedly (CARE & UN Women, 2020). Populations in need of greater protection because of the Tropical Storms included women, girls, adolescents, people with disabilities, LGBTQI+ people, and indigenous and Afro-descendent populations. Damage to health facilities, disruption of transport and increase in demands on healthcare resulting in a reduction in access to health services. Up to two million people in Honduras had limited or no access to health services and remained at increased risk of contracting COVID-19. More than a month after the storms hit, several thousand people remained without access to clean water and health services because of damage to essential infrastructure (Project HOPE, 2020). As of 1 December 2020, 400 health facilities had reported damage, of which 120 were reported inoperative. At least 12 health facilities reported damage to their cold chain equipment, which has disrupted the refrigeration of critical medicines and vaccines (Project HOPE, 2020). As a result of cascading hazards, there were high numbers of dengue, malaria and leptospirosis outbreaks due to stagnant water, lack of access to safe drinking water and lack of sanitation due to the floods and mudslides (CEPAL, 2021; IFRC, 2020b). Damage to crops and agricultural production assets and loss of daily wage livelihood options contributing to food insecurity. The agriculture sector was severely impacted, affecting access to income, goods and services. Losses of up to 80 per cent were accounted in the sector due to damages from both Tropical Storms (CLAC, 2020). The agricultural crops with the greatest damage in 2020 were: coffee (49 per cent), banana (27 per cent), plantain (7 per cent) and sugarcane (7 per cent), through a reduction in the area planted, loss of equipment and therefore a reduction in productivity, and loss of harvest (BIDIDB, 2021). Furthermore, damage to fishing boats and equipment affected the livelihood base of indigenous coastal communities – although there is a distinct scarcity in information about sparsely populated indigenous areas (CEPAL, 2021; FEWS NET, 2021; IFRC, 2020b). There are gaps observed in terms of impact. The main humanitarian reports focus on the urbanized Cortés, Atlántida and Santa Bárbara departments and especially the Sula Valley (IFRC, 2020a; OCHA, 5 For an overview of the administrative boundaries in Honduras, please refer to the map in Annex 7. 20 2020b). However, there are major blind-spots in the understanding of impacts – especially in rural areas in terms of livelihood impacts – that would warrant attention for future initiatives in terms of data collection and humanitarian support (CEPAL, 2021). For example, La Mosquitia – home to the Miskito people among others – was severely impacted by storm surge and river flooding induced by Eta, although other departments have also been impacted by severe floods. Colón, Atlántida, Cortés and Santa Bárbara departments, home to Indigenous Peoples (IP) such as the Lenca, Maya Ch’orti and Tolupan have suffered direct impacts, with sometimes entire colonies destroyed (World Bank, 2020). Box 1. The Chamelecón neighbourhood in San Pedro Sula Various news reports share perspectives from people living in the Chamelecón neighbourhood of SPS where, weeks after Eta and Iota hit, streets and houses were still filled with mud and stagnant water(Tiempo, 2020). The neighbourhood is one of the most violent and poorest areas of SPS, with gangs fighting for territorial control (Brekke, 2021). Where there were flood barriers, these were not strong enough to hold back the flood waters. Homes, business and even the local cemetery were flooded during Eta and again during Iota (Amnesty International, 2020). Residents repeatedly mentioned that, in this area no government support came through – only occasional police patrols (Orellana & Arita, 2021; Verza, 2021). As a result of the devastation left by Eta and Iota, and the limited options for reconstruction, many families remain displaced for prolonged periods of time, living on the streets or under bridges (Olson, 2020). Medium- to long-term impacts Prolonged displacement. While many people returned as soon as the floodwater receded, around 7,500 people remained in 181 shelters as of March 2021 – especially in the municipalities of El Progreso, Choloma, San Pedro Sula, La Lima (OCHA, 2022). Prolonged food insecurity. There was a marked trend of increasing food insecurity in Honduras after the 2020 shocks, with a projected 2.6 million people expected to experience elevated levels of acute food insecurity by August 2022(IPC, 2022). Besides the damage and losses incurred during Eta and Iota, several other factors drove this increase, including job and income losses during COVID-19 and global food prices as well as the food insecurity conditions that existed beforehand because of multiple shocks experienced by the population. Marco-economic trends. By 2020, Honduras’ GDP suffered a drastic loss of 9 per cent, affecting the Government budget along with the private sector. However, it saw an economic rebound the year after, even overcoming the loss of 2020. Growth came from increased remittances and exports, but poverty levels did not improve as quickly compared to the GDP – signalling those economic gains did not reach all the population (World Bank, 2021b). The country presented a growth of 11.9 per cent in 2021 and it is expected to maintain the trend by 3.1 per cent in 2022 and 3.6 per cent in 2023 (World Bank, 2022c). In comparison with other Central American countries, the economic rebound after 2020 has been slower in Honduras (Elliott et al., 2022). Some key economic sectors such as coffee exports have felt the cascading impacts (Reuters, 2021). Migration. News reports from Honduras after Eta and Iota detail stories of families deciding to attempt to migrate to the United States after losing their houses and belongings (Olson, 2020), although it is difficult to discern the relative influence of Eta and Iota compared to the COVID-19 pandemic, violence and other ongoing crises. In 2021 and 2022, caravans have returned to the pre- COVID-19 dynamics, albeit reduced in volume because of stricter migration regulations and controls. Early 2021 marked the highest increase in recorded encounters with Honduran migrants at the US- Mexico border, and 168,546 encounters were registered over the whole of 2021 (U.S. Customs and 21 Border Protection and Mexican Unit for Immigration Policy, Registration, and Person Identity, as cited in Corson & Hallock, 2021; France 24, 2022; Telam, 2022). Those starting from Honduras or passing through are experiencing difficulties accessing food, drinking water, shelter, health, protection and other essential services – especially in municipalities close to busy crossing points (DG ECHO & ACAPS, 2022). However, it must be noted that migration in Honduras is a multi-causal phenomenon that, currently, combines the impact of COVID-19 and socioeconomic vulnerability (as will be further explored in Section 2.2). Impacts on fragility, conflict and violence. In 2020–2021, the long-term reduction in homicides (indicative of gang violence intensity) plateaued in Honduras (Figure 2). In 2020, a lower homicide rate was observed than in 2019 (38.6 compared to 44.7 per 100,000 inhabitants). However, analysis of monthly homicide rates suggests that in the aftermath of Eta and Iota violence intensified compared to earlier in 2020 (Figure 3). Following this trend, total number of homicides increased in 2020–2021. As will be discussed in Section 2, homicides reduced during the COVID-19 lockdown. The strict lockdown started in March 2020 and reopening started in June 2020, although curfew and movement restrictions remained in place until 2021. Homicides were already on the rise before the Tropical Storms hit Honduras. The year 2021 saw a return to the overall high-intensity violence the country experienced before the lockdown-related reduction, with several peaks in homicides in the first months of 2021. Honduras remains one of the countries in the region with the highest homicide and femicide rates (CEPAL, 2021). Although changes in homicides are reported, the levels remain high and continue to have a significant effect in the daily lives of ordinary citizens. Figure 4 shows the number of homicide incidents per municipality in 2021 (SEPOL, 2022) – homicide rates per municipality in 2021 can be found in Annex 4. Overall, homicide incidents were the most frequent in the capital city and north-west of Honduras in 2021, similar to 2020 and earlier years. Of the departments, Cortés and its municipalities were the most impacted areas. (Child) recruitment and business extortion are major drivers of internal displacement in Honduras, and those in temporary shelter (after Eta and Iota) were extra-vulnerable to these and other violence-related impacts (CARE & UN Women, 2021). Figure 2. Timeline of violence, represented by total number of homicides per year in Honduras (Graph 22 by the authors, data from SEPOL, 2022) Figure 3. Count of monthly homicide incidents nationwide between January 2020 and December 2021. (Graph by the authors, data from SEPOL, 2022) Figure 4. Number of homicides in Honduras (2021), map produced by the authors. Data: SEPOL (2022). Politics. At the beginning of 2022, there was a change of administration at all levels of government. Xiomara Castro was elected president as the most voted governor in the history of Honduras. This is a factor to consider since the events analysed in this paper occurred in the middle of the electoral campaign and the recovery coincided with the electoral process and political transition. It is beyond the scope of this paper to evaluate how the Tropical Storms influenced institutional renewal. 23 In summary, the high number of people (temporarily) displaced and extensive infrastructure damage, in addition to impacts on (rural) livelihoods have jointly contributed to an extremely high number of affected populations. Unsafe conditions that resulted in these impacts include the large population, number of buildings and other assets located in flood- and landslide-prone areas, as well as the impacts of Iota in an area still suffering from Eta’s floods and landslides. Long-term implications of the Tropical Storms, combined with other major shocks and stressors present in Honduras at the time are still being felt. The following sections will explore why, where and how these impacts occurred by analysing how accumulated risks and compounding events converged in Honduras. Section 2: Drivers of risk To understand the development of the complex humanitarian crisis that ensued from Tropical Storms Eta and Iota the next sections will i). outline the cascading and compounding hazard interactions in November 2020 when Eta and Iota struck Honduras; ii). identify how natural hazards and violence in the two years preceding the crisis affected exposure, vulnerability and coping capacity to Tropical Storms Eta and Iota and how they contributed to compounding effects in the aftermath; and iii). explore the root causes of selected major impacts observed during the crisis. 2.1. Cascading and compounding hazards Focusing on the north-west of Honduras, several hazards interacted over time and space to produce the severe impacts observed in November 2020 (Figure 5). Although already much decreased in intensity by the time it reached Honduras, Tropical Storm Eta brought heavy rainfall and high winds (see Annex 1 for a detailed description). As a result of the prolonged heavy rainfall and pre-existing physical conditions, several major rivers flooded and landslides occurred (see Annex 1 for flood extent maps). As was the case during other hurricane events (SPS, 2017), the Chamelecón and Ulúa Rivers were the main source of flooding. Tropical Storm Iota compounded the ongoing flooding and landslide impacts of Eta. Although rainfall intensity during Iota was far less compared to Eta, the floodwaters had not yet receded and river and reservoir levels were still very high (IFRC, 2021). As a result, the flooded areas expanded drastically. In the days before the landfall of Eta, there had been a lot of rain in the north and northwest of Honduras due to a cold front (Álvarez, 2020). Furthermore, the soil was already highly saturated due to the record-breaking number of storms that affected Honduras in the months before; and, while the dams had previously been empty due to the prolonged drought, now reservoirs were already nearing their maximum capacity (Conde, 2021). The heavy rainfall during fall of 2020 was partly driven by the La Niña phase of the El Niño–Southern Oscillation (ENSO) phenomenon, typically associated with above average rainfall and storm activity (CENAOS, n.d.). The timing of Eta and Iota at the end of the hurricane season meant that the physical conditions already increased the likelihood of both riverine flooding and landslides. Heavy rainfall directly damaged crops and resulted in several cascading impacts that caused widespread damage. 24 Several KIs indicated that the flooding of the Ulúa and Chamelecón Rivers was more intense and rapid than expected by communities “The problem with everything: flooding based on their risk knowledge (Humanitarian worker 3-6; Civil was so fast once the storm arrived, and society organization). Ongoing deforestation in the upstream the northern area of SPS has a system of areas in Cortés and Santa Bárbara is a major driver of increased canals and embankments that help runoff and sedimentation of the rivers, which in turn caused river prevent flooding in certain areas, but levels to rise more quickly than expected (Academic 2-4; the government had not repaired all Humanitarian worker 9). The Bark Beetle plague in 2016 has been those damaged barriers for years. So mentioned as a further driver of deforestation, along with the the disaster was because the water got clearing of forests for agriculture (FAO, 2019). Furthermore, into all the holes in the embankments ongoing urbanization in both the upstream areas and the low-lying and suddenly flooded everything” areas minimized the infiltration of rainwater and increased runoff. (Humanitarian worker 6) The high flow intensity and rapid increase in water levels as a result was above the design standard of the flood barriers in San Pedro Sula (CEPAL, 2021). Adding to this, several KIs mention that some areas in San Pedro Sula, especially the Chamelecón neighbourhood (see Box 1), suffered from severe flood impacts because the flood barriers breached (Government 1; Academic 1 and 4). This has been blamed on both the intensity of the flooding as well as on the poor maintenance of the flood defences in an area where the soil is known to be very loose and to erode quickly (Academic 4; Humanitarian worker 6 and 7). The rapid and intense onset of floods, breaching and overtopping of flood defences and the presence of a large population and infrastructure in the floodplain led to the mortality and damage to infrastructure observed along the Ulúa and Chamelecón Rivers especially (CEPAL, 2021). Figure 5. Hazard interaction during the dual tropical cyclone crisis in November 2020 against a background of violence in Honduras. As shown in Figures 5 and 6, besides the cascading of the Tropical Storms’ heavy rainfall to floods and landslides, several events before the Tropical Storms deeply affected the vulnerability, coping capacity and exposure of communities in Honduras. For health specifically, the ultimate increase in demand 25 for healthcare and loss of access for those already in need, resulted from a combination of both the floods and landslides related to Eta and Iota as well as the lingering effects of COVID-19 and dengue among other challenges (Figure 6). However, the compounding of the various hazards over time and space only partly explains the extensive losses and damages as a result of Eta and Iota. The following section will further explore the dynamic pressures on vulnerability and exposure. Figure 6. Summary of cascading and compounding dynamics for health impacts. 2.2. Dynamic pressures on vulnerability and exposure The following section explores the accumulation of risk over time and space in the years before Tropical Storms Eta and Iota. First, conflict and migration are explored, followed by a historical disaster analysis to understand other major drivers of risk. Per event, this section explores how this influenced exposure, coping capacity and vulnerability (socio-political, physical, economic, environmental), followed by an analysis of the geospatial co-occurrence of these disasters combined with violence and pre-existing vulnerability. A context of violence and fragility restricts movement and increases economic vulnerability Honduras is affected deeply by fragility and violence. Results indicate that on an individual level, exposure to violence drives displacement to flood- and landslide-prone marginal areas, increases individual social and economic vulnerability and influences choices when faced with disasters such as the ones triggered by Eta and Iota. During the Tropical Storms, the high insecurity, especially in the suburbs of San Pedro Sula, severely limited the dissemination of storm warnings, reduced the willingness to evacuate and limited access for disaster response – as will be further explored in Section 3. Figure 4 shows the homicide incidents rates per municipality in 2021 (SEPOL, 2022). 26 Gang violence is widespread in and around urban areas, and violence related to narco trafficking “Generally, the people who are most affected [read: by particularly affects rural areas (Human Rights Watch the hurricanes] are people who live in those areas, high- (HRW), 2021). Gangs exercise territorial control over risk areas in the Sula Valley, for example. People who, neighbourhoods and, particularly the Mara due to their condition of poverty, cannot live in other Salvatrucha (MS-13) and the 18th Street gang areas, so there are already identified areas where gangs (Barrio 18), are considered responsible for proliferate, but people cannot get out of that cycle since Honduras’ very high murder rate and are infamous they do not have the conditions to move to a better place for extortion and drug peddling(HRW, 2021). In to live.” (Humanitarian worker 3) Honduras, gangs or maras are small groups of young individuals who extort money from local businesses in exchange for protection. Getting caught in the crossfire of conflicts over territory, drug routes and forced recruitment are major risks for civilians. Monthly homicide data shows a sharp decline in homicides nationwide between January and April 2020 as well as the rapid increase to pre-lockdown levels and above from May 2020 onwards (Figure 2). Analysis shows that in the departments of the north-west, violence levels remained high during the months before the Tropical Storms hit Honduras. In 2020, the restrictions on movement and economic activity to curb the COVID-19 pandemic altered the economic situation of the maras drastically. The temporary closure of non-essential businesses meant less opportunity for money laundering, while transportation limitations hindered extortion rackets. As extortion is a main source of income for maras in Honduras their income dropped dramatically – by around 80 per cent (Vazquez, 2020). Drug trafficking was also obstructed by travel restrictions, which caused difficulties for both users and traffickers; the global decrease in recreational drug use; and the limited supply of chemicals required for drug production (Dalby, 2020; Eligh, 2020). As this economic pressure increased on maras, inter-gang violence increased across Honduras – compounded by a conflict between maras and the state due to several incidents in prisons (Armed Conflict Location & Event Data Project (ACLED), 2020; Academic 4). Some gangs turned to the extortion of transport businesses, while others threatened to collect retroactive extortion payments when restrictions were lifted – severely increasing the economic vulnerability of affected groups across Central America (Dalby, 2020). Once restrictions were in fact lifted, in the summer of 2020, the maras turned to various methods to make up for lost extortion payments. Some demanded back-payment for lost funds, while others stepped up drug peddling, armed robbery and contraband (International Crisis Group, 2020). Despite changes in the economic pressure exercised by gangs, annual data on homicide rates per 100,000 population show that 2020 homicides were much lower compared to earlier years – from 2017–2019 the figure varied from 41.4 to 44.7, and it decreased to 38.6 in 2020. This is also true at the local level – in all municipalities of Honduras data from the World Bank indicates that in 2019– 2020, most municipalities saw a decrease in violence. Only 18 municipalities – of which five are in the Lempira Department, three are in Olancho and three in Copán – saw an increase in homicide rates in 2020 (SEPOL, 2022). These reductions should be perceived as highly relative. It is important to note that despite the positive developments in terms of homicides observed in recent years, Honduras remains amongst the ten most dangerous countries in the world. For example, San Pedro Sula municipality in 2020 had a homicide rate of 20.97 per 100.000 population – which was the lowest rate since 2015, where London had a rate of 1.32 in the same year. Therefore, the impact of this improvement on the daily lives of citizens is considered to be limited. 27 Ultimately, criminality can affect every single aspect of people’s lives. Being a victim of a homicide or losing relatives this way naturally has not only a personal impact but also an economic one, since there is one less person to contribute to the often already irregular and low income of the poorest households. Criminality similarly impacts the choices people make about where to live, forcing some to leave their homes along with their social and economic connections or alter their daily activities – e.g., by choosing a lower-paid job to avoid having to travel at night. Finally, criminality even influences how people react to disasters such as tropical storms, as will be explored in Section 3. Migration crisis driving urbanization and vulnerability Migration in Honduras is a major factor affecting socioeconomic “Migration has two faces, violence dynamics in the country. In 2018 migrant caravans to the United causes migration, but migration helps States began from Honduras. That same year there were four fuel cycles of violence in the sense that caravans between October and December. The caravans were it dismantles homes and communities.” stopped during 2020 because of the pandemic restrictions, but (Humanitarian worker 2) resumed once borders re-opened in 2021. The number of encounters at the US-Mexico border with Hondurans is a measure for the number of migrants, and this increased from 135,261 in 2018 to 347,285 in 2019. As a result of COVID-19-related policies, the number of migrants dropped considerably to 78,121 in 2020 (U.S. Customs and Border Protection and Mexican Unit for Immigration Policy, Registration, and Person Identity, as cited in Corson & Hallock, 2021). Besides livelihood pressures, rural–urban migration and intra-urban displacement is driven by violence (Humanitarian worker 3; WFP, 2017). The areas where poor migrants settle are some of the most insecure areas around San Pedro Sula, where access to government support and social safety nets is very limited (ICRC, 2017). Cortés and Choloma – areas severely affected by the Tropical Storms – are the main destination for internally displaced populations, and households are often displaced within the same department and even municipality, often settling in high-risk areas (CIPPDV, 2015). Within San Pedro Sula, displacement because of urban gang violence has been recognized as a major problem influencing economic and social vulnerability (ICRC, 2017). Being at odds with the dominant gang, refusing to pay “protection fees” or acting out of line can be life threatening, forcing many to flee the neighbourhood. The departments of Cortés, Choloma and Yoro also receive many migrants who have been returned by Mexico. In 2019, there was a record-breaking high number of returnees (109,185 people) (CEPAL, 2021). With few alternatives, and often already in deep debt to smugglers from the migration attempt, many (temporarily) stay in the cheaper neighbourhoods of San Pedro Sula (Humanitarian worker 9; WFP, 2017). While the number of returnees reduced drastically during COVID-19, those who did arrive in Honduras could not move out of the Sula Valley area, as public transport was disrupted (Academic 1 and 4). This resulted in a higher number of people living in unsafe conditions, with limited economic means and social networks to cope with shocks. “San Pedro Sula, wherever it is, as I say, the industrial capital of the country, they go there or they go to the United States or Mexico, but a large number of people stays in San Pedro Sula and they begin to settle in high-risk areas, living in poor conditions, so there was a large number of people who had newly arrived in the area.” (Government 1) Besides national migration and displacement, San Pedro Sula at the time also housed a large population of international migrants (mainly from Haiti and other Latin American countries) who were 28 on their way to the United States, as borders had been partially re-opened in the months before. International migrants are among the most vulnerable populations, with limited access to basic services and precarious living conditions that are typically condensed in high-risk areas or on the streets (Pulte Institute, 2020). At the time of Eta and Iota, these groups had limited access to information as well as solidarity networks and were among the most economically fragile. Figure 7 summarizes the main migration dynamics and captures some of the linked disaster events, which will be further explored below – both in relation to violence and migration as well as independent drivers of vulnerability and exposure. Figure 7. Migration dynamics in Honduras and interaction with major disaster events in 2018–2020. Political context In addition the complex socioeconomic framework in Honduras, several other events contextualized the compound risk scenario during the impact of Eta and Iota. The year 2017 ended with the re- election of President Hernández, under accusations of fraud, and exploded into violent social protests that left at least 23 people dead and led to more than 1,300 detentions (Daugaard, 2019; Human Rights Watch, 2019). In 2018, the Attorney General of New York charged the former Honduran Congressman and brother of the current president of Honduras with “conspiring to import cocaine into the United States and related firearms offenses” (US Department of Justice, 2018). In September 2019, the former First Lady of Honduras, Rosa Elena de Lobo, was sentenced to 58 years in prison for “undue appropriation of funds and fraud” (Reuters, 2019). In 2018–2019, President Hernández implemented a Police Reform, which resulted in almost half of over 13,500 police officers being dishonourably discharged (Human Rights Watch, 2020). Trust in the Government was affected by allegations to corruption and high violence rates, according to key informants. 29 Natural hazard-related disasters in the previous years An historic disaster analysis was conducted for the three years prior to the two Tropical Storms impacting Honduras (Figure 8). A total of seven major natural hazard-related disasters were identified for the period January 2018 to November 2020: hydro-meteorological (hurricanes, floods, droughts, landslides) and biological (COVID-19 and dengue), which took place in a context deeply affected by violence and migration. A list detailing the events and sources can be found in Annex 7. Below, we discuss the natural hazard-related disasters that had consequences for the impacts of Eta and Iota, to this end we excluded the flooding and landslides in 2018 and generalized the disaster event of Tropical Storm Amanda to the ‘2020 hurricane season’. Figure 8. Natural-hazard related disaster events timeline in Honduras from January 2018 to December 2020 and annual homicide count in Honduras between December 2017 and December 2020 (data source: SEPOL, 2022). Drought – July 2018 (2014–2019) In 2014–2019, Honduras experienced severe drought in the “Dry Corridor” region of the country triggered by the El Niño phenomenon. The Government of Honduras declared a state of emergency in affected areas from 15 August to 31 December 2018 (FAO, 2018). Following prolonged droughts over a five-year period, the Central American Regional Climate Outlook Forum forecasted below average cumulated precipitation from El Niño in Honduras in early 2018, which further aggravated the climate conditions (FEWS NET, 2019). These accumulated effects resulted in the worst drought in ten years in 2018, diminishing the capacity for livelihood recovery in the country (WFP, 2018). The main impacts were on subsistence agriculture, disrupting the main harvest season (Primera), as well as loss of income for daily wage laborers. While market prices remained average, food insecurity became acute in the affected areas (FEWS NET, 2018). The drought conditions resulted in shortages of basic grains, water scarcity and lack of 30 pasture for livestock, with reported losses of 72 per cent for corn and 75 per cent for bean, as well as significant losses in livestock according to the country’s Ministry of Agriculture and Livestock (ReliefWeb, 2019). About 65,500 families required emergency food assistance in 34 municipalities at the start of the emergency (Government of Honduras as cited in Masters, 2019). It is estimated that the drought impacted a total of 360,000 people in 2019 alone in the departments of Choluteca, Comayagua, Copán, El Paraíso, Francisco Morazán, Intibucá, La Paz, Lempira, Ocotepeque, Olancho, Valle and Yoro. The uncommonly long duration and intensity of the drought that affected the Dry Corridor in the years before Tropical Storms Eta and Iota, including parts of Honduras, has been partly attributed to anthropogenic climate change: temperature rises and other climatic changes have resulted in a four- fold increase in the probability of such drought events occurring (Pascale et al., 2021). Climate change is driving a drying trend in Honduras, where observations already indicate that droughts have become more frequent, intense and affect a larger geographical area (I. T. Stewart et al., 2022). These trends also contribute to long-term challenges in soil quality, water saturation and deforestation through an increased likelihood of forest fires, which influence the long-term vulnerability of communities dependent on these natural resources. The prolonged drought conditions in the Dry Corridor of Honduras served as a push factor for rural– urban migration, which resulted in the expansion of informal settlements in the Sula Valley – increasing the population exposed to the floods and landslides that occurred during Eta/Iota. Rural population vulnerability had accumulated over time, as the multiple-year drought had affected the food sources of subsistence farmers, and had limited income-generating opportunities for daily labourers working in the agricultural sector, especially in coffee production (FAO & WFP, 2020). The World Food Programme (WFP) has found that the main push factor for migrating away from agriculture-based rural communities in Honduras is food insecurity (WFP, 2017). While some people directly attempt international migration towards the United States, many first move to urban centres in Honduras – especially San Pedro Sula and Tegucigalpa. San Pedro Sula is the main industrial hub in the region, and the drought meant that farmers moved (seasonally) into this urban area to find work in the industry and services sectors (Moloney, 2019). Furthermore, San Pedro Sula is one of the main departure points for international migrants aiming to reach the United States as well as home to returnees coming from the US–Mexico border (Humanitarian worker 2). With few resources available and limited opportunities to generate income because of COVID-19 restrictions, many stayed with family or in the cheapest housing possible, which is typically poorly constructed and located on marginal lands on the outskirts of the city (Academic 4). These informal settlements had expanded into flood zones over the previous years (Humanitarian Worker 9) – the very areas that were severely affected by riverine floods and landslides as Eta/Iota impacted the low-lying Sula Valley. “There were still restrictions due to the pandemic, and we were in a period of drought, and that was when people concentrated in the main population centres, and not in an orderly manner. We saw that, for example, San Pedro Sula collapsed in terms of its capacity to receive population that comes from the interior of the country as a result of the drought” (Government 1) Dengue – 2019 Honduras experienced the worst dengue outbreak in its history in 2019. An Epidemiological Alert was released in Honduras and a State of Emergency was declared in response on 14 June 2019 (Potter, 2019). Dengue fever is a disease principally found in urban and semi-urban areas of tropical and 31 subtropical regions, spread by mosquitoes that breed in stagnant water. This disease is endemic in Honduras and outbreaks occur every four to five years (MSF, 2019). The combination of prolonged drought, high temperatures and a subsequent rainy season in early 2019 had led to above average mosquito proliferation (IFRC, 2019). The political unrest at the time of the outbreak as well as violence worsened the epidemic by restricting access to some of the most severely afflicted neighbourhoods, limiting awareness-raising and preventive measures such as fumigation (Potter, 2019). According to PAHO/WHO, a total of 132,143 cases were reported during the 2019 outbreak with approximately 30 per cent infected with dengue haemorrhagic fever – a more severe form of the disease – leading to 180 fatalities, with children making up nearly half of the reported deaths (PAHO, 2019). The dengue outbreak directly affected the health system, with impacts still felt at the time when Eta and Iota struck. According to the Honduran Ministry of Health, over 75 per cent of all cases originated from north-western Honduras with the most affected areas in Atlántida, Cortés, the Central District, Olancho, Santa Bárbara, San Pedro Sula and Yoro – the areas where floods and landslides during Eta/Iota also caused the most damage to health infrastructure (OCHA, 2019). During the outbreak, out of the 32 public hospitals in Honduras, 26 were over capacity with patients needing to be treated for dengue (IFRC, 2020a). The nation’s poorly supplied and understaffed medical facilities lacked the capacity to adequately handle even the nation’s normal demands, much less an epidemic of historic proportions requiring a fast, highly organized response (Semple, 2019). While an Emergency Appeal was launched and significant aid commitments were made, doctors, nurses and organizations involved in the response lamented the lack of preventive measures taken and raised concerns over the worryingly high insecticide resistance of mosquitos as well as corruption across the health system (MSF, 2019). In addition to the acute strain on the health system due to COVID-19, the dengue outbreak in 2019 and the ongoing endemic conditions complicated both the diagnosis of COVID-19 patients and contributed to a highly overstretched system during November 2020. Furthermore, the dengue outbreak had impacted the north-west especially hard in the previous year, “What COVID did allow us to understand is that a health and the limited prevention measures pre-exposed emergency is a systemic risk, that is, the risk is not linear, the area to a resurgence of dengue (IFRC, 2020a). it cascades, it impacts everything, so you look at a KIs mention that the burden of dengue, combined pandemic that suddenly has an economic crisis, of with COVID-19 and gastro-intestinal diseases limited schools, transport, ministers, then it makes you see that the access to quality care during and after the we can no longer think of things in isolation and that the Tropical Storms – especially for those populations risks are really interconnected, they are difficult to dependent on the public health system foresee.” (Humanitarian worker 1) (Humanitarian worker 4; Civil society organization 4). COVID-19 – 2020–present Both the direct health impacts and the economic fallout from the COVID-19 pandemic severely impacted economic vulnerability and people’s coping capacity for Eta and Iota. Honduras recorded the first cases of COVID-19 in March 2020. In total, 121,827 COVID-19 cases and 3,130 fatalities were recorded in 2020. Within a few days of the first confirmed cases, the country went into strict lockdown, shutting down all non-essential businesses and restricting the movement of citizens who were only permitted to leave home once every two weeks. Both the health implications of COVID-19 in Honduras, where the health system was over-stretched already, and the economic consequences severely impacted the population. 32 Over 70 per cent of Hondurans, especially women, work in the informal sector, where unemployment, poverty, homelessness and food insecurity increased during the COVID-19 pandemic (Díaz-Bonilla et al., 2021). The significant economic impact of COVID-19 saw an increase in migrant flows from rural to urban areas (DTM, 2021). Food insecurity also increased for population groups that were previously not affected by food insecurity, especially in the services and industry sectors because of the lack of economic opportunities as well as widespread loss of income and employment. Due to the inadequate health system, which suffered from chronic underfunding and corruption, as well as the ongoing severe dengue epidemic, many patients were unable to receive care (ASJ, 2020). Nonetheless, health facilities were improved slightly with the funding received for the COVID-19 response (Humanitarian worker 4). The ongoing stress on the health system and emergency management coordination structures contributed to the lack of access to healthcare during and after the Tropical Storms (see Figure 4). The socioeconomic impacts and severe strain on public spending drastically reduced the coping capacity of the population affected by Tropical Storms Eta and Iota. Rates of gender-based violence, especially domestic abuse, increased during the pandemic (CARE & UN Women, 2020). The increased poverty and unemployment left women and girls at an even greater risk of multiple manifestations of violence, exploitation and human trafficking with the highest femicide rate in Latin America during 2020 and 2021 (UNSDG, 2021). People in the informal markets, small and micro entrepreneurs, women in precarious employment conditions, and historically excluded groups such as indigenous and afro-descendants, are identified as the most vulnerable to the impacts of COVID-19 (UNDP, 2021). During the dual Tropical Storm crisis, these dynamics compounded the precarious shelter conditions, resulting in high GBV rates and protection concerns. The government response to the COVID-19 crisis was widely critiqued, and the measures imposed severely affected small enterprises and the informal sector. Protests against the COVID-19 movement restrictions turned violent. Successive abuses by the security forces during protests against the COVID- 19 lockdown imposed by the national Government were denounced during 2020 (Human Rights Watch, 2021). Furthermore, an audit revealed millions of US dollars of misappropriation in the medical supplies acquisition to combat COVID-19 by the Government (Human Rights Watch, 2021). Tropical Storm Amanda and the 2020 hurricane season Tropical Storm Amanda hit the western parts of Honduras on 30–31 May 2020, impacting 249 families and leading to five fatalities. Amanda was one storm out of the 30 named storms and hurricanes during the 2020 Atlantic hurricane season – an extraordinary active year (Thompson & Montañez, 2020). Where natural hazards and violence collide Both the long-term impacts of hazards in 2018 and 2019 as well as the various disasters that occurred in 2020 before Eta and Iota have influenced the crisis. However, not all hazards affected the different regions in Honduras equally. The geospatial analysis visualizes which areas were affected by the most disasters, against a backdrop of ongoing violence, vulnerability and (lack of) coping capacity (LOCC). This was a layered approach, of which the results are summarized in Figure 9. The findings help focus the retrospective analysis on the areas where multiple stressors likely affected the population. 33 Figure 9: Municipalities impacted by compounding hazards in Honduras (Map by the authors, data sources summarized in Annex 7). The municipalities of Yoro and Catacamas were the areas most affected by the historical disasters, combined with LOCC and high vulnerability as per the INFORM index. These areas did see impacts from Eta and Iota (CEPAL, 2021), although far less people were affected compared to other departments. The population density in Yoro and Catacamas is very low (for an overview of population density, please see Annex 5). Therefore, in the next step, population density was considered explicitly to understand where the largest affected population was located. When focusing on the areas affected by the various disaster events and violence, corrected for population density, findings from the hotspot map indicate that the areas most affected by floods during the Eta/Iota crisis (the municipalities Choloma, El Progreso, La Lima, Puerto Cortés, San Manuel, San Pedro Sula and Villanueva – see Annex 3) are also the areas that have historically been affected by a high number of disasters and a very high incidence of violence. Furthermore, for La Ceiba, Omoa and Tela municipalities (in orange on the map), it should be noted that there are urbanized areas with high population density, but the total population of the municipality is relatively low (see Figure 10). The municipalities with the two largest cities of Honduras, Distrito Central (Tegucigalpa) and San Pedro Sula, had by far the highest incidence of violence in 2018–2020 as indicated in blue (for a map of conflict incidence only, please see Annex 6). The following map shows the analysed hot spot municipalities, overlayed with population density. This map highlights how strongly population density varies across Honduras. 34 Figure 10: Municipalities impacted by compounding hazards and population density in Honduras (Map by the authors, data sources listed in Annex 7) 2.3. Root causes of risk To understand the construction of risk beyond the major shocks discussed in Section 2.2., this section explores the root causes of the dynamic pressures that produced the unsafe conditions during Eta and Iota. The root causes of vulnerability involve social and economic structures, such as the characteristics of power, wealth and resources distribution, as well as ideologies and historical heritage (i.e., war and post-war fragility, militarism). Such root causes may change, albeit slowly (Marchezini & Wisner, 2017). The section summarizes the key findings from several impact-specific PAR analyses (Annex 4) in one overarching PAR in Figure 11. “Unsafe conditions” are the direct causes of why people and their assets were in “harm's way” (Wisner et al. 2014; Oliver-Smith et al. 2016). In the case of Eta and Iota, the high number of people and their assets as well as public infrastructure and animals located in the areas affected by floods and landslides are the main unsafe condition that resulted in high displacement and losses. However, disaster management-related factors (further detailed in Section 3), such as the lack of adequate warnings for the flood and landslides, limited willingness to evacuate, inadequate organization of shelter and ongoing violence also contributed to the outcomes observed in Section 1. Dynamic pressures summarized in the PAR graph in Figure 11 and the more detailed impact-specific PARs in Annex 4 are a result of the major shocks covered in Section 2.2 as well as root causes. These root causes will be further explored in the following section, structured along four types of vulnerability: socio-political, economic, physical and environmental. 35 Figure 11: Pressure and release framework for general disasters during Eta and Iota in Honduras Social and political vulnerability Decentralized governance with limited local capacity to perform tasks At the municipal level, the Law and Regulations of Municipalities determine that it is their responsibility to prevent and respond to disaster events and see to the development of plans and strategies for disaster risk reduction. Municipalities also have the responsibility for (urban) planning and have an important role in the management of land-use in rural and urban areas. However, there are major differences between the financial and human resources capacity between municipalities in Honduras (Development actor 1 and 2). In several municipalities, this results in a vacuum where regulations and plans at national level are not implemented or maintained at local level. Urban planning guidelines are available in some more urbanized areas, but interviewees mentioned these are not fully followed – resulting in settlement in high-risk areas, poor building quality and the limited extension of public services (water, sanitation and hygiene (WASH), education, electricity) (Development actor 1 and 2; Academic 1 and 2). For DRM, this situation forces municipalities to turn to the highest level of government, which is the national Government represented in COPECO. This results in a lack of disaster risk management planning outside of major urbanized areas, e.g., no emergency plans, limited functioning of local governance structures such as emergency committees and early warning systems (see Section 3). During Eta and Iota, this led to differences in warning and evacuation capacity, and especially the organization of shelter, which specifically is a responsibility of the municipality. Many municipalities did not have all the capacity to identify and prepare locations in advance nor to organize according to international standards, resulting in lack of shelter spaces, overcrowding, understaffing and lack of access to essential services such as WASH (CEPAL, 2021). Monitoring of shelter conditions is also lacking. Corruption limiting investment in public systems, regulation and law enforcement Corruption permeates through all aspects of Honduran public life, affecting the maintenance of critical infrastructure, access to essential services and the overall development and recovery after major shocks. Recently, Hernández, the President of Honduras at the time that Eta and Iota struck, has been 36 trialled for corruption in the United States (US Embassy Tegucigalpa, 2022). Corruption at the various levels of Government hinders the implementation of national and regional regulations, for example, on urban planning and building codes – putting more people at risk as settlement in high-risk areas in sub-standard housing is tolerated. In rural areas, deforestation and the expansion of large agri- businesses on communal lands is also enabled through a network of corruption linked to the narco trade (McSweeney et al., 2018; Tellman et al., 2020; USAID, 2017). Corruption also influences disaster risk reduction measures. In San Pedro Sula, some of the breaches of flood defences and damage to bridges was attributed to a history of diversion of infrastructure maintenance funds (Humanitarian worker 8; Government 1). Furthermore, COPECO itself had been implicated in a corruption scandal before the Tropical Storms, resulting in the ousting of many high- level officials (Corson & Hallock, 2021). The person left in charge of COPECO after the scandal at the time of the Tropical Storms admitted later that he did not have the experience to deal with the crisis (Corson & Hallock, 2021). The lack of transparency complicated the fundraising and distribution of aid funds before, during and after Eta and Iota for several organizations (Development worker 1). These events limited the trust of citizens in the warnings and recommended early actions and have eroded the trust of citizens – especially youth – in the Government (Corson & Hallock, 2021). Lastly, the healthcare system is one of the essential services that has been chronically affected by corruption, as well as insufficient funding and substandard management (see Health-specific PAR in Annex 4). A long history of political corruption has pushed the healthcare system to the edge of collapse, with reports in 2014–2015 that 49 per cent of the public health budget was diverted from health, resulting in an estimated 2,800 related deaths (LAWG, 2020). The chronic lack of funding has led to limited resources with overcrowded facilities and dwindling basic protective equipment even prior to the COVID-19 outbreak (Borgen Project, 2020). Violence and drug trafficking As well as being a risk dynamic pressure (as described in section 2.2), the protracted violence in Honduras, intertwined with the prevalence of drug trafficking, is rooted in the social dynamics of the country. Dealing with violence is part of day-to-day life for most Hondurans, including recruitment, general security concerns and the risk of dispossession (Human Rights Watch, 2021). In urban areas affected by gang violence, coverage by government aid can be limited and protection fees leveraged by criminal groups can severely increase economic vulnerability. In rural areas, the lack of land security and limited protection makes people vulnerable to land grabs and other forms of coercion – making violence one of the main reasons for rural–urban and intra-urban displacement (WFP, 2017; ICRC, 2017). There is a strong link between corruption and organized crime in Honduras, with the President, his brother, and other political and business elite facing trial or accused in the US justice system of collaborating with narco-related organized crime (Corson and Hallock, 2021). Inequality and exclusion Pronounced gender inequalities and the exclusion of indigenous peoples (about 9 per cent of the population according to the 2013 census) from decision-making, capital and social services puts these groups especially at risk. For example, up to one million people, largely from IPs, were excluded from food support (the “bolsa solidaria”) during the 2019 drought and food insecurity crisis (Criterio, 2020). The high number of women employed (and underemployed) in the informal industry resulted in a high loss of jobs during the first year of the COVID-19 pandemic, resulting in a “feminization” of poverty (CARE & UN Women, 2021). The high out-migration of breadwinners has resulted in a high number of single-parent households (the majority female-headed) and children under the care of the elderly, 37 with families often in debt and unable to finance the migration attempt. These families are more often affected by food insecurity, dependence on aid and are extra vulnerable to shocks (WFP, 2017). Economic vulnerability Poverty and inequality Multidimensional poverty and high inequality have affected vulnerability and coping capacity throughout Honduras historically, and were amplified by the COVID-19 pandemic and the associated economic recession. In 2019, 14.8 per cent of the population lived below the international poverty line and 25.2 per cent of Hondurans lived in extreme poverty – about 2.2. million people (World Bank, 2021a, 2021b). While extreme poverty is mostly concentrated in rural areas, inequality is especially pronounced in urban areas. Honduras had a high GINI index of 0.482, signalling high income inequality (World Bank, 2022d). Both rural and low-income urban areas suffer from a lack of access to health, sanitation and education as well as limited livelihood opportunities, severely limiting their coping capacity to shocks such as the dual Tropical Storms in 2020. Insecure land tenure and land-use change Long-term lack of access to safe, affordable and productive land for agriculture and construction is a structural issue. As a result, many Hondurans settle in rented houses and plots in marginal areas. Land tenure in Honduras is unsecure because of limited inter-institutional coordination, poor conflict- resolution mechanisms, limited cadastral and legal information and limited recognition of communal lands – resulting in a constant threat of invasion or eviction (World Bank, 2017). Land is often not properly titled, resulting in ambiguous ownership and this leaves especially those with limited means for litigation vulnerable. This affects both rural and urban areas, where the majority of land is owned by a small minority (Recio, 2015). In urban areas, this lack of affordable land is resulting in development at the periphery (Recio, 2015). In rural areas, minifundistas (small-scale farmers) are among the most vulnerable – the insecurity also limits their access to credit and therefore their ability to invest in increasing productivity. Furthermore, IPs often live on communal lands without clear titles and have been historically affected by encroachment and dispossession by other farmers and large- scale corporations (Herlihy & Tappan, 2019). Legislation in Honduras also supports land-use change from forests (essential for many IPs’ livelihoods and food) to coffee crops and the construction of buildings as legal proof of ownership (Recio, 2015) – affecting IPs and accelerating deforestation with implications for long-term disaster risk reduction and sensitivity to drought and heavy rainfall. High fiscal deficit limiting public spending The international debt of Honduras was a major theme in the 2020/2021 presidential elections, partly as 2020 saw a rapid increase in the Government financial deficit and up to 30 per cent of the Government budget went towards debt payments (Palencia, 2021). During this year, debt service payments also increased (Elliott et al., 2022). Combined, these macro-financial pressures severely limited the Government’s ability to invest in the public health system, education, maintenance of critical infrastructure and support for municipal DRM (Elliott et al., 2022; Civil society organization 1) – although several shock mitigation measures were taken (World Bank, 2021a). Import/export imbalance and reliance on remittances While Honduras’ GDP has recovered since 2020, the growth has been driven by exports and remittances (World Bank, 2021a) resulting in increasing inequality and producing limited benefits for 38 the poor. This skewed recovery finds its roots in the increasing reliance on export, especially agricultural cash crops to the United States, and this has made Honduras vulnerable to macro- economic shocks. For example, during the COVID-19 pandemic, coffee exports and industrial trade decreased markedly – putting many informal unskilled labourers out of work and increasing economic vulnerability (World Bank, 2021a). High out-of-pocket expenditures for health On average, Central America spends 14 per cent of GDP on healthcare, whereas Honduras allocates only 8.5 per cent of its GDP expenditure to health (CDC, 2021). At present, the main sources of health financing in Honduras are 54.7 per cent Government, 34.4 per cent household out of pocket, and 8.2 per cent external sources (Carmenate-Milián et al., 2017). These pre-dispose low-income families to catastrophic health expenditures and make access to healthcare sensitive to economic shocks. Access to healthcare for families in Honduras is also determined by poverty level, socioeconomic status and whether they live in a rural or urban environment. In rural environments, healthcare is much harder to access despite efforts to improve these conditions through the national health model implemented in 2015 (Brigada, 2020). Although the Ministry of Health provides care to 88.3 per cent of the total population, a majority of the services are in the most developed cities, making healthcare inaccessible for rural and indigenous populations (CDC, 2021). Undiversified and informal labour market Food insecurity in Honduras is rooted in the systemic unequal distribution of income and resources, rapid urbanization, vulnerability to extreme climatic events, undiversified and largely informal labour market, inequitable access to land, and insufficient food production (WFP, 2022). For those not dependent on agriculture, the informality of the labour market and lack of diversification, combined with ongoing migration from rural agricultural areas towards the districts of Cortés and Atlántida mean increasing competition for employment. In 2020, over 13 per cent of adults lost their jobs during the compounding crises, mainly in the informal sector (World Bank, 2021b). This predisposed many to economic hardship during the COVID-19 pandemic, which severely affected the incomes of those especially unskilled and underemployed in the industry, services and tourism industries, as well as labour opportunities in coffee and banana plantations. A full analysis of the root causes driving these vulnerabilities in food security specifically can be found in Annex 5. Environmental vulnerability Expansion of large-scale agriculture and deforestation Degradation – especially soil erosion and deforestation – contribute to environmental vulnerability, causing the more rapid onset and increasing the extent of flooding during Eta and Iota; pre-disposing communities in marginalized areas to landslides – as mentioned in Section 2.1. Furthermore, the areas with high impacts from landslides had been previously affected by deforestation (CEPAL, 2021). Rapid deforestation as a result of agricultural expansion, limited control on logging, as well as the increasing incidence of wildfires, are contributing to environmental degradation. Intensive monoculture farming is also considered to accelerate land erosion (FAO, 2019). Deforestation in the flood-affected areas was also accelerated through the Bark Beetle plague in 2016 (FAO, 2019). While erosion is a natural process, the expansion of large-scale agriculture combined with the drying trend observed across Honduras are accelerating land degradation and soil erosion, which in turn decrease agriculture production and change runoff and flooding patterns (FEWS NET, 2019). 39 Limited ability to invest in agricultural productivity Degradation of lands, especially already marginalized lands farmed by small-holder farmers, contributed to the flooding and landslides – but also predisposed these farmers to loss of production and lower production. This was aggravated by the multi-year drought and resulted in a rapid increase in food insecurity as reductions in yield reduced options to generate income and crops for consumption (beans, maize). The previously discussed insecurity of land tenure and limited options for obtaining credit limit smallholder farmers’ ability to invest in technologies or equipment to deal with such shocks. The planting of marginal lands and uncontrolled practices increase soil degradation further (FEWS NET, n.d.). A lack of waste management and drainage contributed to environmental vulnerability to the floods, but also increased the transmission and proliferation of dengue, waterborne diseases and overall poor shelter conditions. Physical vulnerability Poor housing quality, overcrowding and settlement in high-risk areas remain major drivers of physical vulnerability (CEPAL, 2021). In “Well, what I wanted to add is the anecdote of the August 2019 the municipality of San Pedro Sula woman from the neighbourhood we visited, it was published guidelines to prevent settlement close really sudden when the dam collapsed and to waterways and on steep slopes, in response everything was flooded. So, of course, there were to the rapid increase of the population in the alerts, but also particular cases that caused certain past years (slopes >25 per cent, in floodplains areas to be suddenly affected. Well, it was not etc). However, a lack of enforcement of this expected that a dam or two would break in San regulation has resulted in the ongoing expansion Pedro Sula.” (Humanitarian worker 2) of settlements in these areas. Poor urban planning practices are known to be one of the main factors underlying vulnerability and exposure. On the former, for example, segregated communities with limited access to jobs and public services result in lower economic development. On the latter, the development of settlements in risk areas due to lack of adequately planned areas to settle expose the most vulnerable people to several different hazards. Informal and unplanned urban developments are common throughout Honduras. In the Sula Valley, for instance, a study from Habitat for Humanity (Hábitat para la Humanidad Honduras, 2019) has identified 133 informal settlements housing approximately 163,000 inhabitants (more than 11 per cent of the total population of the region). From these, about 42 per cent are sitting in risk areas, which would already put them in danger regardless of their vulnerability conditions. Yet, more than 55 per cent of the houses in these communities are built from leftover construction materials – far below adequate standards to withstand even mid-scale disasters. Lastly, the physical characteristics of the valleys and riverbanks most affected by flooding also contributed to physical vulnerability and unsafe conditions. The hilly terrain in the north and west of Honduras means human settlements are located close to rivers and expand up the steep slopes, and in the Sula Valley multiple rivers converge – increasing the flood exposure. Furthermore, the loose soil in the valley means that flood defences and landslide prevention measures need to be maintained regularly, as erosion is fast (Academic 3). 40 Section 3: Early warning signals, response and predictability of impacts The following section explores what early warning signs were apparent ahead of the compound crisis, and to what extent were they acted upon by relevant authorities, focusing on Eta and Iota and their impacts as well as the potential for anticipating impacts of a similar nature in the future. 3.1 Early warning and response for Eta and Iota This section is structured along the four building blocks of effective, people-centred early warning systems (World Meteorological Organization, 2018, p. 2): I. disaster-risk knowledge based on the systematic collection of data and disaster risk assessments; II. detection, monitoring, analysis and forecasting of the hazards and consequences; III. dissemination and communication, by an official source, of authoritative, timely, accurate and actionable warnings and associated information on likelihood and impact; IV. preparedness at all levels to respond to the warnings received. I. Disaster risk knowledge Information about disaster risk (including exposure, vulnerability and coping capacity) is the foundation of early warning systems (WMO, 2018). In Honduras, Hurricanes Fifi (1974) and Mitch (1998) were major reference points for risk information. In the aftermath of Mitch, several international agencies supported the Honduran Government and knowledge institutes like the Universidad Nacional Autónoma de Honduras (UNAH) to systematically collect disaster information(Cárdenas, 2012). As part of SINAGER, COPECO and UNAH have developed a municipal level INFORM index tool, which summarizes crisis risk (hazard and exposure, vulnerability and lack of coping capacity) for all 298 municipalities (INFORM, 2021). The main gap is that this is only available at municipal level (HRC, 2019). There is also a geonode for Honduras, which serves as a repository for geospatial data (http://geonode.copeco.gob.hn/). The areas most affected by flooding during Eta and Iota – the Ulúa, Chamelecón and Choluteca basins – were understood to be highly exposed to floods (Academic 4; HRC, 2021), with several informants expressing that this is common knowledge. Nonetheless, KIs also mentioned that flood and landslide risk perception was low among communities living in high-risk areas (Government 1; Humanitarian worker 4). The main reason for this low awareness according to KIs was the relatively high number of youth and adolescents in Honduras who had not experienced a severe hurricane yet like Hurricane Mitch (1998) and were, therefore, less aware of the potential impacts. II. Forecasts The Centro de Estudios Atmosféricos, Oceanográficos y Sísmicos (CENAOS) (National Centre for Atmospheric, Oceanographic and Seismic Studies) within COPECO is responsible for monitoring and forecasting weather-related conditions (COPECO, 2016). For hurricanes, CENAOS receives support for monitoring and forecasting from international agencies, especially the National Hurricane Center (NHC) of the US National Oceanic and Atmospheric Administration (NOAA) (Pasch et al., 2021; S. R. Stewart, 2021). CENAOS uses the forecasts from the NHC and adds a rainfall forecast based on the Weather Research and Forecasting model. For hurricanes, there are no defined impact levels – although CENAOS–COPECO do use a colour-coded system for warnings. There is currently no understanding of the different impact levels of hurricanes, and there is a disconnect between hurricane forecasts and flood forecasts (Cárdenas, 2012; Government of Honduras & World Bank, 41 2019; HRC, 2019). Key informants mention that this lack of clarity on impacts associated with issued weather forecasts is the main reason for incorrect low-risk perception among the population, and leaves people uncertain about the best preventive actions to take in their locality. At the time (and still), flood forecasting depended on the monitoring of river water levels only, and rainfall forecasts were not included in flood monitoring. Early warning systems in Honduras for floods and landslides are typically organized at the community scale and are not connected to weather forecasts, with limited records and little data going from the local level to national level (Cárdenas, 2012; HRC, 2019). For floods, this involves monitoring the amount of rain over a particular period, or the river levels based on painted scales on bridges, carried out by volunteers who alert neighbours and the municipality (Civil society organization 1). For Eta, several KIs mentioned that the flood onset was more rapid and intense in the Ulúa and Chamelecón basins, defying the “common sense” of those monitoring the river levels and catching communities by surprise (Humanitarian worker 6). As discussed in Section 2.1, the sedimentation of these rivers resulted in the more rapid rising of the floodwaters. At the national, regional and municipal levels there are major organizational challenges that limit effective monitoring and forecasting. CENAOS was, and still is, a small team of forecasters and support staff with major challenges in operational capacity (Government of Honduras & World Bank, 2019). This includes a lack of trained staff that can interpret and process weather data, a lack of finances to expand services and restraints in the organization that limits CENAOS’ ability to fulfil its mandate within COPECO. Furthermore, monitoring infrastructure is outdated and poorly maintained (Humanitarian worker 9). At the national level, staff are monitoring forecasts; however, the connection with local systems is often missing (Government 1; Humanitarian worker 1, 3 & 9). As part of the national COPECO, there are regional COPECO offices which are in contact with a ‘forecast centre’ that usually sits within the municipal mayor’s office (Cárdenas, 2012; COPECO, 2016). It was mentioned by several KIs that this communication chain is often too slow, and people at municipal level do not have the means or the training to work with forecasts (Government 1; Humanitarian Worker 1, 3 & 9). An inventory study from 2012 highlighted that, in many cases, the responsibility for forecasting is not well established, there is a neglect of weather warning systems by the municipality, and the monitoring systems depend on paper records kept by volunteers (Cárdenas, 2012). Tropical Storm Eta and Iota forecasts Based on the forecast analysis and warning review (full analysis can be found in Annex 2), the following key points help to quantify the available warnings of the Tropical Storms: • Forecasts for Eta were issued in advance, but the rapid intensification of the storm before it hit Nicaragua and its slow movement over Honduras, as well as the change in trajectory, were not forecast accurately. Around 30 hours before landfall in Nicaragua, forecasts did show these dynamics. The change in track of Eta meant that the north and north-west of Honduras were affected by the storm for longer than expected in the initial forecasts. • Tropical Storm Eta not only brought strong winds, but it also resulted in very heavy precipitation in the north and north-west of Honduras, triggering flooding and landslides especially in the Copán, Cortés and Yoro departments. Heavy rainfall was forecast very well by many agencies, including CENAOS, and the global European Centre for Medium-Range Weather Forecasts (ECMWF) model had good performance seven days in advance of the rainfall peak. • Iota was forecast exceptionally well, with a longer lead time compared to Eta. Tropical Storm Eta damaged many of the gauges and windspeed measurement equipment, making analysis difficult. 42 It should be noted that there were no forecasts available of the potential impacts of Eta and Iota – i.e., the losses and damages that could result from the strong winds, heavy rainfall, riverine flooding and landslides. Furthermore, there were no national or international flood alerts shared, although the international Global Flood Awareness System (GloFAS) model was publicly available for the relevant rivers (Kopp & Meriam, 2020). “Achieving confidence in the forecasts that they send us is a challenge in the communities, because the damage that a flood can cause is not measured and communicated” (Civil society organization 2) “I don't think anyone outside of COPECO “[About Hurricane Eta] If now if you ask me as a knows the difference between a green meteorology professional, it was certain that I was going and yellow alert, what's more, I don't to enter, there was no doubt about that, the doubt was know, I can make a good estimate, but I regarding the territory in the trajectory of the honestly don't know.” (Civil society hurricane.” Academic 1 organization 1) Figure 12. Communications and dissemination of warnings: Tropical Storms Eta and Iota 2020 (Source: compiled by authors based on NHC 2020, 2021 and COPECO news reports) III. Dissemination and communication of warnings6 Figure 12 shows the timeline of the national level communications of official warnings and advisories as well as communications from NOAA, to understand what information the authorities, communities and other agencies had available for anticipatory action. 6 COPECO website was offline during this analysis, which hindered the analysis of the warnings. 43 As mentioned, the forecasts for Eta had a relatively short lead-time due to the erratic track and rapid intensification into a tropical storm. COPECO shared NHC hurricane forecasts in a timely manner, yet was perceived as too slow in sharing the rainfall warnings, which were available several days earlier (the warnings for Cortés and Yoro were only shared on the 3 November by COPECO, while rainfall forecasts were already available on the 29 October). As a result, there was a spatial disconnect between the areas where initial hurricane alerts were issued and the regions that eventually saw the most impact (from flooding). This disconnect in the moment meant that people perceived the north- west and northern region of Honduras as far away from the storm track and, therefore, relatively safe, while this area had the highest flood risk, and the early onset of heavy rain and associated flooding caught people by surprise in the north and north-west. Furthermore, people did not feel at risk. This was because of the spatial scale of the warnings at district level as well as the absence of impact information for specific hazards in specific areas (heavy rainfall, floods, landslides). In addition, Hondurans receive multiple hurricane warnings every year, and the general public assumed Eta and Iota would be similar to previous storms (Civil society organization 4; Government 1). Particularly in 2020, there had already been many alarms and alerts due to COVID-19 and the intense hurricane season. This severely limited the actionability of the alerts, as people were generally unaware of the precautionary steps and confidence associated with the risk levels. For Iota, COPECO shared alerts even ahead of the first advisories by NOAA – five days in advance (See Figure 12). The organization also shared an advisory calling for early action four days in advance – a step that was missing for Eta. While the messages shared by COPECO and the NHC did not contain any impact-based warnings beyond the district level, some news agencies did share information on the neighbourhoods at risk of flooding. For example, in San Pedro Sula, Tiempo – a Honduran daily newspaper – shared information on the areas at risk and instructions on preparatory actions i.e., packing an emergency kit and preparing a family evacuation plan (Tiempo, 2020). However, many communities were still cut-off from communications networks and transport as a result of Eta at the moment warnings for Iota were shared. Dissemination and organization CODELs may have different communication methods, such as driving around with loudspeakers, sharing information through social media or sounding a siren. However, several interviewees mentioned that this system is only operational in some areas and there are large differences between municipalities and their capacity (trained staff, protocols, finances) to organize themselves and share warnings (Government 1; Civil society organization 1; Academic 1). Furthermore, these government entities are often distrusted and may not have access particularly to gang-controlled areas (Government 1; Civil society organization 2). Besides official government channels, Hondurans rely on sharing of information through social media, especially WhatsApp (groups) (STC). Despite the economic situation in Honduras, there were 79 cell phones per 100 inhabitants available at the time (World Bank, 2022a). The parts of the population that do not have access to social media or the internet receive information through community members and neighbours as well as the radio. However, because of the heavy rainfall and floods around San Pedro Sula, the large volume of emergency calls caused the telecommunications network to collapse (Government 1). In areas where communities were physically cut-off from the outside due to the Eta floods, WhatsApp was the only lifeline through which to access information and communicate needs to humanitarian organizations (Humanitarian worker 6; Civil society organization 1). As official channels were out of service, and WhatsApp communication was primarily maintained 44 between community members and civil society organizations, the Government and community could not communicate directly. IV. Early actions and response In the departments with red alerts by COPECO before Eta and Iota, the main early actions carried out were mandatory evacuations and the sharing of warning messages; but, for the most part, action was only taken when the impacts were already materializing (flooding, landslides). Plans and protocols for early action and response Some municipalities (in the Cortés and Yoro departments) have Emergency Plans and Risk Management Plans (developed under the World Bank-funded Proyecto Gestión De Riesgos De Desastres project (Disaster Risk Management project) (COPECO, n.d.), that helped the local authorities take action. For municipalities in other affected departments (e.g., Copán, Santa Bárbara) the protocols and capacity to respond to warning messages shared by COPECO was limited. San Pedro Sula – one of the areas that was most affected by the heavy rainfall and flooding caused by Eta and Iota – had an Emergency Plan in place since 2017 (COPECO, 2021) containing information on the areas exposed to flooding, landslides and wildfires. It also outlines evacuation routes, decision- making processes, protocols depending on the level of the warning issued (green, yellow, red) and suggestions for early action on all the hazards identified (for an example list, see Annex 2). The Emergency Plan, therefore, contains crucial information, although several important gaps were identified too: the risk analysis focuses on exposure and not on vulnerability (e.g., quality of houses, coping capacity of the population); there is no planning for compounding/multi-hazard situations; and there is limited attention given to early actions to take in the window of opportunity between the first warnings and the actual event. Evacuation is the main approach, but the Plan mentions that shelter locations are only identified and equipped when the warning is issued. Eta demonstrated that the lead time for warnings can be very short; which, in this case, did limit the options for early action. Several of the measures mentioned in the Emergency Plan would be considered preparedness actions. However, preparedness for an event of the magnitude of Eta and Iota was missing. During the Eta/Iota crisis, San Pedro Sula did not have shelters organized in advance, attributed partly to the rapid increase in its population and the limited access to certain neighbourhoods by government bodies due to gang violence. This was a crisis magnifier (Academic 1). Examples of effective early actions in Honduras Local media reported on several early actions (prepositioning, mobilization of people) across Honduras for essential food, shelter and non-food items along with the mobilization of civil defence to support evacuations (Brackett, 2020; el Diario, 2020; Redacción Web, 2020; Valladares, 2022). When asked about examples of areas that did take effective early action for Eta and Iota, several informants mentioned that the highly exposed Islas de Bahía performed very well because of a sound understanding of risks, availability of shelter, knowledge of emergency protocols and, overall, a higher capacity to deal with tropical storms. Furthermore, the islands were not exposed to riverine floods. Interestingly, the absence of violence on the island was quoted as the main reason for the higher capacity and willingness to evacuate (Academic 1, Government 1). Furthermore, the San Manuel Cortés, Omoa and Tela municipalities were said to have evacuated effectively and organized their shelters well. Overall, the Cortés department performed well compared to Yoro and Santa Bárbara, attributed to the fact that they did do some work on developing Emergency Plans and community based EWS (effective to varying degrees). The Honduras Red Cross also mentioned that it had some successful interventions in the south of Honduras, mainly because it actively monitored the forecasts 45 and prepared shelters, evacuated people and coordinated with the municipality in a timely manner (Civil society organization 3). However, in general, challenges in the preparedness and response during the crisis outnumbered good practice, in retrospect. Challenges in early action and response Late evacuation and limited shelter options Given the size and intensity of hurricane events, evacuation from at-risk areas is often one of the main early actions to prevent loss of human life. Evacuation was left to the last moment because of multiple reasons. First, there was confusion about the risk levels used by COPECO and the late announcement of the warnings for the Sula Valley meant that people did not understand the severity of the impending tropical storms. The communications from the Government did instruct people to evacuate but did not share information on how and where to go (Government 1; Civil society organization 1; Academic 1). The coordination of the evacuation was also mentioned as a major bottleneck (Humanitarian worker 7; Academic 1) – shelters were not identified and ready, while municipal and COPECO staff were not available (see Morazánica holiday, below). Ultimately, many of the evacuation orders were carried out by military and police forces. For the hardest hit urban areas, reports suggest that there were not enough shelter spaces available (Olson, 2020). After Eta, many locations were already full, and when Iota struck people were reluctant to move to overcrowded shelters. Furthermore, a critical consideration is that the floods in the Sula Valley affected some of the poorest and most violent areas, and the fear of being robbed or the inability to return made people unwilling to leave their houses in advance (Humanitarian worker 3 and 6; Government 1). Coordination with heads of churches and community committees was essential in these areas (especially Baracoa, Celeo Gonzáles, Cortés, El Progreso, La Lima, La Planeta, Rivera Hernández and southern San Pedro Sula/Choloma) to even allow response teams to enter (Spring & Ceja, 2020). Coordination of action Most organizations included in this research mainly discussed the response to Eta. When Iota was announced, organizations started early mobilization and took measures as they already had an operation on the ground. There were challenges in the humanitarian response, as staff in Honduras was already stretched after a long emergency situation due to the COVID-19 pandemic (Humanitarian worker 3) and international funding support was far below the target. Several KIs mention that the small civil society organizations and the church had an important role in reaching the most vulnerable communities and conducting needs assessments. The ability of COPECO and other government agencies to communicate relied mainly on radio and telephone lines, which were affected by power cuts and damage to communication infrastructure. Communication and the coordination of the disaster response between the Government and international organizations was disjointed and, among national actors, the various operation centres at various levels of government did not always coordinate effectively (Humanitarian worker 7; Government 1). The different instructions and decisions from national to local levels resulted in the duplication of effort (Government 1). Flooding of the airport in the Sula Valley complicated supply chains further (Humanitarian worker 7). Furthermore, there was a lack of organization and coordination of roles and responsibilities among the main national government entities. Later, the head of COPECO at the time admitted that he and his organization where wholly unprepared to manage a disaster of the scale of Eta/Iota (Corson and Hallock, 2021). The diverging decisions from 46 the leadership also limited the implementation of many of the existing protocols and plans (Humanitarian worker 3). Semana Morazánica The weekend before Eta stormed over Honduras, there was the long-awaited national holiday Semana Morazánica, which had been postponed due to the COVID-19 lockdown. The holiday was a major driver of the overall low preparedness of Honduras for Eta, for several reasons: • Movement of a large part of the population. Hondurans who could afford it were on holiday or visiting family from 27 November and, therefore, away when the first warnings of Tropical Storm Eta were issued (2 November) (Civil society organization 4) • The holiday was long anticipated, and the north of Honduras is a popular destination (because of the beaches; many people also had family there). At first, the north was also perceived as less at risk from the Tropical Storms, as it was far away from the storm track and heavy rainfall initially forecast, while the associated flood risk for the north was not communicated effectively (Humanitarian worker 9). Combined, these reasons motivated many to disregard the warnings of the Tropical Storm and prioritize their holiday plans. • Some informants suggest that there was political pressure to prioritize the holiday over warnings for Eta, delaying evacuations and other important prepositioning (Humanitarian worker 7, Government 1, Academic 1) • Decision-makers were out of the office at the time warnings for Eta came in, delaying the preparedness and response mechanisms (Government 1). COPECO staff and civil defence were spread out across the country to set up checkpoints to control holiday-related traffic (COPECO, 2020). • Those people affected by Eta while on holiday had less knowledge of safe evacuation procedures, locations of shelter and limited access to local support structures (Civil society organization 3). 3.2. Predictability of impacts Building on the analysis of the drivers of risk before the Tropical Storms along with the cascading interaction of hazards during Eta and Iota as well as the assessment of early warning capacity for the Tropical Storms at the time, the following section explores if and how the disaster impacts could have been anticipated and what would be required to develop such a system in the future. Based on the impact pathways defined in Section 1, Table 1, below, outlines findings and evidence on the predictability of short-term impacts at the time (November 2020) using available data and tools. It also outlines a potential predictive system that could be explored to enable risk-informed decision- making to anticipate potential crises, hence reducing risks and preparing for an effective disaster response in the future. It is important to invest in approaches to anticipate the impacts of events like Eta and Iota to be prepared in the future. For the hurricane events, changes in global climate dynamics are projected to result in more storms of this intensity that are also slow moving and bring intense precipitation (IPCC, 2021; Kossin et al., 2017; Patricola & Wehner, 2018). While no attribution study has been conducted for Eta and Iota specifically, their characteristics do fit the observed and projected trends of more intense storms in the Central America region (Kossin et al., 2017). Projections suggest that while the 47 number of storms on average may remain stable, there will be a higher number of very severe storms (IPCC, 2021). Eta and Iota could serve as a reference point for potential impacts of such severe events. Ability to anticipate crisis impacts in November 2020 At the time of the crisis, there was no multi-hazard early warning system in place and this severely limited the ability of actors in the country to foresee the humanitarian crisis in November 2020 and beyond. The hazard interactions were not properly recorded in an integrated, standardized and coherent manner, and socio-economic vulnerability information was not part of the DRM system. When Tropical Storm Iota hit first, there was anecdotal understanding of which areas were more at risk, yet there was no organized system to support COPECO, different ministries, local communities and humanitarian and development actors to make more effective decisions in advance of the shocks. Furthermore, the focus on response rather than prevention as well as the structural challenges in the fiscal management of COPECO and the municipalities, limited long-term preparedness (Civil society organization 3; Humanitarian worker 5;8). Due to limitations in the effectiveness of regional COPECO offices, disaster risk management responsibility often lies with the municipalities. However, in many areas, the municipalities do not have the trained dedicated staff, resources and other materials necessary to carry out early action plans for potential disasters (Academic 1). The overall assessment to determine at what extent disaster impacts could have been predicted emphasizes a disconnect among the information management systems and disaster risk management measures across key thematic areas, such as food security and agriculture, health, migration, and violence. An overview of available information sources for the various hazards, risk factors and drivers of risk is provided in Annex 5. An overall conclusion is that these information sources were not utilized to their full potential and monitoring systems were incomplete, resulting in crucial knowledge gaps at the time of the crisis in November 2020. For example, as described in Section 2.2, Honduras faced a severe humanitarian crisis due to drought, induced by ENSO, in 2014– 2019. This crisis heightened food insecurity levels, inducing malnutrition. However, those facing the direct and indirect impacts of the drought were not tracked and supported to cope with the subsequent hazards. This accumulation of risks was compounded by the disruption of the health system due to the dengue epidemic in 2018/9; the agriculture damage due to the floods in 2018 and early 2020; and the extreme impacts across all sectors of COVID-19. Understanding population movement within Honduras at this time along with the key vulnerabilities and lack of coping capacities of rural families would have helped to better understand the impacts of the Tropical Storms. Rural areas were hit hard by the multi-year drought; saw high out-migration and economic fragility in the years before Eta and Iota; and daily wage labourers especially had been affected by the economic impacts of COVID-19. Nonetheless, reporting on Eta and Iota focuses mainly on urban areas, leaving gaps in our understanding of the dynamics in rural Honduras. For the effective anticipation of crises of a similar nature, improvements in risk data availability will be crucial. In November 2020 – e.g., for flooding and landslides – there was exposure information available in urban areas; although its accessibility, spatial scale and accuracy were often limiting factors. It is unclear how often available exposure data is updated, and risk assessments lack information on vulnerability. According to various KIs, the available information was also not used in the preparedness and response coordination at the time, signalling the need for institutionalization of available science. In the case of the Sula Valley, the breaching of critical flood defences also resulted in unexpected impacts and limited coping capacity, and this confirms the need for regular field assessments to understand whether maintenance is carried out and regulations are followed, e.g., 48 whether flood barriers have been maintained, drainage networks are free from potential blockages, and urban planning is followed. There also remain major information gaps in terms of vulnerability, with a need for integration on monitoring systems. For example, rural and sparsely populated areas are poorly mapped compared to urbanized regions and there is little information on the experiences of indigenous communities. The location and vulnerabilities of international migrants staying in Honduras temporarily was a particular gap identified. Another crucial gap was the lack of integration of monitoring systems for migration and violence from the natural hazards-related DRM system. The risk analysis showed that these drivers had a major influence on the size of the exposed population as well as their vulnerability, coping capacity and disaster response (physical access of actors to these areas; access to aid, finances and information). Anticipating similar crises in the future From a technical perspective, impact-based forecasting could be feasible. The understanding of the accumulation of risk over space and time could be possible by uniting various information sources; although, admittedly, with higher confidence in mapped and analysed urban areas. However, there are several bottlenecks in terms of risk information management, including limitations on granularity of data and the limited centralization of information. This would likely limit impact predictions to neighbourhood level or larger, and the accuracy would be highly dependent on regular updates of risk information given the dynamic violence and migration conditions in Honduras. In November 2020, decision-support tools and DRM systems were not in place to monitor dynamic risks, take anticipatory action, or act effectively post-disaster, thereby increasing primary and secondary impacts (e.g., diarrhoeal diseases, COVID-19 because of poor shelter, drowning because of late evacuations, loss of livelihoods because of lack of warnings). The table 1, below, focuses on the exposure and vulnerability indicators that could have been part of an impact-based forecasting system which can be used by disaster risks managers to identify who and what is more likely to be impacted in the face of certain hazards. In the case of the 2020 Tropical Storms, the probability of exceeding 300 millimetres of accumulated rainfall was 70 per cent with a lead-time of 7–10 days. It is important to note that landslide predictability in Honduras requires further research. The potential actions to prevent impacts are based on early actions and DRR interventions from other contexts, tailored to the Honduras context. 49 Table 1. Overview of indicators and early actions for specific impacts, based on the findings for Tropical Storms Eta and Iota Impact What happened at the time? Potential indicators: Exposure Potential indicators: Vulnerability What can be done with information to prevent, reduce impacts and prepare for disaster response in the future? Death by Combination of flood, landslide, - Population located in - DRM: municipalities Tailor awareness messages through various drowning and mudslides, compounded by floodplains or close to flood /neighbourhoods without DRM communication channels, with guidance specific to or burying the violence dynamics in the barriers, on steep and unstable plans at-risk neighbourhoods on how to avoid death by country that limited, delayed or slopes and a large distance from - Violence-related: Areas with high drowning and burying through timely evacuation and prevented evacuation drainage networks levels of homicides; Government’s avoiding high-risk situations (crossing rivers, getting -Population with lack of access to lack of access to neighbourhoods close to damaged electrical infrastructure) Prediction feasibility: Predicting evacuation sites - Household composition: specific mortality would have been - Areas with a high presence of percentage of people with Prevention of floods and landslides through difficult, although there was maras disabilities, female- and elderly- structural mitigation measures and .nature based detailed exposure data available - Location of high number of headed households solutions, such as reinforcing high risk slopes with a for the urban areas migrants (international/national) mechanism to strengthen soil. Investment in the capacities of local and national organizations such as the Honduras Red Cross. Increasing community based DRR teams and the volunteers network, increasing the use of tools such as OpenStreetMap. Building on traditional and indigenous knowledge to reduce risks. Loss and Combination of floods, landslides, - Location, number and building - Coping capacity: Poverty levels, Preventing and reducing risks: damage of and mudslides. Status of drainage characteristics (materials, malnutrition levels, people with -Short-term measures to reduce infrastructure homes infrastructure, building and number of storeys etc.) of cardiovascular and respiratory damage: strengthen flood defences, sandbag, and resulting in material quality, location near infrastructure (houses, WASH, diseases, number of elderly- or pumping of water displacement streams or on steep slopes schools, roads, etc) in floodplains female-headed households, -Targeted evacuation efforts of most at-risk areas, under influences impact of hazard or close to flood barriers and renting versus owning property, including coordination with civil society distance to drainage networks organizations, the Honduras Red Cross, churches and 50 precarious Prediction feasibility: -Number of evacuation shelter people with disabilities, people protection agencies to access most violent areas to conditions Infrastructure factors could all spaces per population, checks on with large families prevent late evacuation, overcrowding of shelters have been predicted, although adherence to Sphere standards - Violence-related: Incidence of and secondary impacts spatial granularity was limited. for shelter gender-based violence; incidence -Long-term measures: avoid damage to houses and Breaching of flood defences would - Previous communicable disease of homicides; number of other infrastructure, reinforcing building codes, risks have been difficult to predict incidence, especially for dengue, unaccompanied minors in shelter informed zoning, controls on investment in public without in-depth monitoring zika, chikungunya; areas with locations; Government’s lack of works, strengthening local coordination mechanisms high incidence of COVID-19 access to neighbourhood - Areas with high presence of - Health-related: see below Protecting families and individuals in high-risk maras locations: -Protect personal documents, essential non-food items, seek higher/stable ground and identify shelter options ahead of time - Implement long-term interventions addressing the willingness to evacuate as well as conditions in shelter locations Preparing for evacuation: - Ensure personal preparedness i.e., packing medical kits, ensuring access to water and sanitation in evacuation places, enabling social distancing as required, mask-wearing other biosafety measures - Ensure food access puts particular emphasis on those with existing food insecurity Damage to Compounded by ongoing COVID- -Number and construction -Access to healthcare, e.g., -Target health warnings to at-risk groups in affected health 19 and dengue outbreaks. characteristics of health number of doctors and nurses areas, including instructions on keeping medicines facilities, Inadequate shelter conditions infrastructure in at-risk areas available per capita, insurance safe, preventive behaviours, and nearest accessible disruption of (lack of WASH facilities, (low-lying, steep slope, far from coverage, out-of-pocket clinics transport and overcrowding) drainage networks) expenditure, population living in -Target humanitarian support to most at-risk increase in -Number of access routes to poverty, distance to nearest neighbourhoods with highest needs and lowest demands on Prediction feasibility: The needs at health facilities; distance of facility before and after the access to healthcare; advance coordination needed healthcare the time could not have been personnel to health facility impact -Stock-up on essential medicines and protective resulting in a predicted, given the lack of -Location of previous -Facility information on medicine equipment; protect medicine stocks, cold chains and reduction in communicable disease stocks, personnel etc. 51 access to predictive tools available for incidence, especially for dengue, -Data on population in need of equipment from floodwaters to prevent loss of health dengue and gastro-intestinal zika, chikungunya in regular care, e.g., non- medical supplies and equipment services diseases. A general increase could neighbourhood communicable diseases, pregnant -Set-up of field clinics and hospitals in most at-risk have been expected given the -Health facilities within areas of women, young children, elderly areas; preparing mobile health clinics to ensure available knowledge on conditions. high levels of violence (with -Access to clean water, hygiene, emergency access to medical care. Effectiveness Information on health facility level history of violent incidents) and sanitation in shelters (in case could be impacted by structural issues with might have been available in urban -Areas with high presence of of a compounding dual Tropical corruption, violence and road conditions areas but was limited in rural areas maras Storm event) -Poverty levels, malnutrition levels, people with cardiovascular and respiratory diseases Damage to No agriculture-specific risk -Location crops (coffee, banana, -Size of agricultural land; yield- Prevention: crops, monitoring services or warnings African Palm etc.) exposed to levels in previous seasons - limited short-term options. Ad hoc flood barriers, agricultural shared at the time of impact floods and landslides -Existing food (in)security status early harvesting, storage protection may be feasible production -Location of land areas with -Household-dependency ratio; assets and Prediction feasibility: detailed environmental degradation, socio-economic status; land Mitigating secondary impacts: loss of daily information available on the deforestation, desertification, tenure; existing level of debt; -Activate shock responsive social protection systems wage croplands exposed to flooding and erosion -Labourers’ socioeconomic targeting at-risk agricultural workers livelihood (Sula Valley), granularity not at -Seasonal calendar of crops: conditions (poverty, migration) -Provide food assistance to populations already options farm level. Food insecurity as a when and where crops are -Incidents of violence in facing the compound impacts. Provide cash contributing secondary impact of these planted and harvested agricultural areas assistance to migrant populations facing food to food agricultural losses was anticipated -Location of agriculture facilities, insecurity (particularly those stranded or returning) insecurity by FEWS NET (large geospatial- warehouses, silos etc. -Establish coordination systems to ensure people scale predictions) -Agriculture land presence of living in areas of high violence are reached maras 52 Conclusions and recommendations going forward This study presents evidence of the severe impacts that sequential tropical storms can have as a result of compounding hazards, risks and vulnerabilities. Affecting the same geographical area in close succession, Tropical Storms Eta and Iota resulted in cascading hazards culminating in widespread flooding and landslides, especially affecting the north and north-west of Honduras (Figure ES2). As a result, communities living close to the major rivers in the north, especially in the Sula Valley, saw their neighbourhoods flooded – resulting in drowning, damage and destruction of infrastructure and cutting off many communities by damaging roads and communication networks. Landslides on the hills at the periphery of San Pedro Sula and in the west of Honduras also resulted in widespread damage. Shelters could hardly absorb the large, displaced population after Eta, and the system was overwhelmed when Iota extended the floods further. Overcrowding and loss of essential infrastructure for water, sanitation and health services resulted in rising COVID-19 cases, dengue and gastro-intestinal diseases. Loss of employment in the affected urban areas, along with loss of (subsistence) crops and labour options in rural areas, contributed to rising food insecurity by limiting households’ ability to buy food. The damage to infrastructure and agricultural lands from the floods and landslides resulted in major economic losses – compounding the pre-existing economic recession in Honduras as a result of COVID-19. Areas affected by violence and the economic consequences of COVID-19 were hardest hit by Eta and Iota, partly due to physical conditions but mainly because of the dynamic pressures on vulnerability and exposure that accumulated over time. Furthermore, the dual Tropical Storms compounded the existing economic crisis as a result of COVID-19, political instability and corruption that had already severely increased economic and socio-political vulnerability. Dynamic migration processes in Honduras result in more people in at-risk areas – especially peri-urban areas – negatively affecting disaster preparedness and action. In the years preceding the dual tropical storms, rural-urban migration was fuelled by the multi-year drought and violence, while intra-urban displacement as a result of violence and poverty forced people to live in high risk areas (floodplains and unstable hillsides), especially around San Pedro Sula. The combined result of unsafe buildings, settlement in at-risk areas, poverty, inequality and ineffective governance increases dynamic pressures on unsafe conditions – including a lack of investment in, and maintenance of, disaster risk reduction interventions as well as health and social security systems that may help communities to cope with such hazards. Furthermore, ongoing environmental degradation, especially deforestation, soil erosion and the accumulation of waste – partly because of climatic factors, but also due to Government and economic processes – have been shown to intensify flooding and landslide onset and aggravate impacts. Root causes of vulnerability predispose Honduras to similar crises in the future. In addition to the root causes of vulnerability, unsafe conditions were aggravated by a lack of early action, mainly because of a lack of understanding of potential impacts, limited dissemination of warnings and lack of protocols, plans and finances to execute anticipatory action. Violence influenced willingness to evacuate and safety in shelters, and lack of trust in the government as well as ineffective coordination between national and local authorities meant early action and the disaster response were slow to materialize. 53 The following section summarizes key recommendations for support to the Government of Honduras and other actors to better anticipate similar crises in the future and to take decisions accordingly. Based on the findings of the compound risk and early warning capacity analysis, this report recommends to: Invest in research of compound risks to guide policy- and decision-making. The evolution of risks over time and across geographies is complex, and impacts from crisis such as Eta and Iota can only be fully understood by analysing accumulation of risk over time and space. A greater investment in continued research, practice and policy (ideally at sub-national scale) will enable the Government of Honduras and its partners to fully understand the dynamic changes in risks, with emphasis on supporting marginalized and poor populations, migrants and those living with violence. Support government, universities and other knowledge institutes to improve data quality, granularity and accessibility for hazards, exposure and vulnerability measures. Data improvement is a cornerstone for an open, safe, effective and harmonized system that can anticipate a crisis. The hydro-meteorological monitoring network across Honduras needs increased investment, while community surveillance systems for biological hazards should be improved. A consistently updated information system is needed for populations living in at-risk areas along with vulnerability profiles of those at risk, details on the movement of people, and changes in the various types of violence – as outlined in Section 3.2. Vulnerability profiles could support the understanding of pre-existing vulnerability and detail potential compounding dynamics, including migration dynamics and push/pull factors that influence vulnerability and exposure (i.e., urbanization and settlements in floodplains, lack of awareness of regional hazard measures, seasonal precarious movement). Social safety net programmes could be utilized to identify, monitor and support at-risk populations. Programme records of households and communities who received support during previous disasters (i.e., 2014 – 2019 drought) can be used to track vulnerability and reach households faster and more effectively. More detailed suggestions for data sources are included in section 3.2. The compounding impacts of the Tropical Storms could have been better anticipated if an integrated risk monitoring system was in place that included existing vulnerability. A system that integrates information on food security, health, violence, housing and migration (in addition to hazard forecasting and monitoring) could have been instrumental in anticipating the impacts of the 2020 crisis. Different governments, departments and non-governmental actors in Honduras have data that is essential to understand who is most at risk. Such a system should be linked to SINAGER and can build on already existing information management systems or can be inspired by frameworks or tools developed by the World Bank, such as InaSAFE in Indonesia, or by Governments, like Nepal’s BIPAD (national disaster information management system). InaSAFE is an impact scenario tool that supports analysis of the potential impacts of natural hazards, which could be combined with information on health, violence, migration and food insecurity in Honduras to understand the accumulation of risk. BIPAD integrates information from the Ministry of Interior, Disaster Management Department, Hydro- Meteorological Department, the Police and others, with opensource data such as OpenStreetMap. The INFORM Index in Honduras is an important step towards the establishment of a harmonized, coordinated and regularly updated multi-hazard risk information management system. To further improve INFORM, the system should bring together information from across COPECO units, related ministries (e.g., Ministry of Health, Environment, Social Welfare), academic institutions (UNAH), humanitarian actors and key actors at the municipal and neighbourhood levels. Additionally, INFORM should integrate a consistent monitoring system into its programme to inform emergency managers of the most updated compound risk information at any given time. 54 Improve modalities to monitor violence as an essential component of disaster risk management. Violence should be considered an integral part of risk monitoring and decision-making for future crises. Urban and rural violence related to maras and drug trafficking groups have been shown to deeply affect all dimensions of risk in the population including exacerbating vulnerabilities, contributing to migration, altering cognitive risks bias, and increasing risky behaviours such as resistance to evacuation. People in areas affected by violence also have less access to government and other humanitarian and development support, predisposing them to a vicious cycle of disaster impacts. Monitoring violence can improve decision-making in times of crisis. A dynamic risk system would inform if, and to what extent, the levels of violence can have repercussions on the capacity to act before a shock and identify subsets of the population that need special protection considerations in times of evacuation. Agencies such as the International Committee of the Red Cross could play a key role on this process. Root causes and structural faults need to be addressed to reduce the risk of disaster. Lastly, this analysis shows the importance of considering compound risk, especially hazard exposure, violence and vulnerability together. For any work on decision-support tools, it remains crucial to holistically address both the drivers of risk and the root causes of vulnerability. Support is required to address systemic under-investment, implement climate change adaptation measures, and improve community safety nets. For example, while the protection of forests as well as nature-based solutions to combat land degradation can enhance the quality of soil to reduce the risk of drought or reinforce areas prone to landslides and floods, these measures need to be accompanied by efforts to limit further deforestation, ensure flood and landslide management infrastructure is maintained and governance systems are in place to make effective decisions. Community knowledge of disasters and understanding of risk can be improved through education programmes in schools and community- level engagement in disaster risk reduction. Providing cash and voucher assistance is widely understood to increase the perception of the severity of imminent disasters and is an effective method to buffer immediate negative outcomes on households. The systemic root causes of vulnerability and exposure are likely to worsen without adequate intervention, especially under a changing climate. Some recommendations are provided in Annex 6. Establish the next generation of impact-based forecasting systems. A major gap in the early warning of Eta and Iota, as well as for earlier events, was the lack of public guidance on the potential disaster impacts and suggested early actions for imminent hazards. Warnings for Eta focused on windspeed and track and overlooked the high risk of flooding and landslides in the north-west of the country. The heavy rainfall was forecast well in advance, even before the tropical storm forecasts, but this window of opportunity was not utilized resulting in late evacuations, poor preparedness and avoidable loss of lives and livelihoods. A comprehensive impact-based forecasting system would help facilitate warnings along with advance support to the most at-risk communities, thereby minimizing the disaster impacts. Enhance coordination protocols and support local capacity-building for disaster response, including the financial capacity to act when required. It is important to strengthen coordination and operational capacity at the district and municipal levels as well as intergovernmental coordination and coordination among humanitarian actors, as there was a disconnect between the national response and local entities during Eta and Iota. The lack of preventative action based on forecasts ahead of the crisis was also attributed to weak coordination between various levels of government, which comes back to the limited capacity at municipal level. In a system where municipalities are the executive arm of the disaster management system, reforms are required for improved communication systems, 55 better trained and motivated staff in key monitoring positions, and funding to maintain these structures and act when needed. Additionally, a platform of coordination between disaster risk management, humanitarian, development, climate and peacebuilding actors is key for climate change adaptation and DRM plans, including anticipatory action. Ensure warnings to key players and the public are timely, convey impacts and provide guidance on preventive actions. Warning communication and dissemination protocols are crucial to reduce the risks of disaster. COPECO, including CENAOS, should be supported to shift towards the use of impact- based warnings. Warning systems should be coupled with campaigns to increase public trust in government entities. Civil society and humanitarian actors such as the Honduras Red Cross can play a role in growing trust in public institutions through their strong community networks. Improvements to communication protocols include clarification of the different colour alerts issued by COPECO, adding impact information into warning maps, sharing advice on potential early actions and establishing communication chains that reach communities. Communication chains can build on local emergency committees, networks of civil society organizations, and the church and other actors. Messaging should be clear, use multiple communication modalities, and communicate the associated likelihood of the event, expected severity of the impacts, and clearly communicate necessary actions (e.g., securing windows, preparing emergency kits, identifying shelter locations, evacuating areas close to rivers). This will be crucial to translate available risk information into tangible results in the prevention of impacts on the ground. References Amnesty International. (2020). Llueve sobre mojado en Honduras: El devastador impacto de los huracanes Eta e Iota. https://www.amnesty.org/es/latest/news/2020/12/devastating-impact- hurricanes-eta-iota-honduras/ Daugaard, A. (2019). Honduras: How a surge of corruption scandals has fueled political crisis. Transparency International. https://voices.transparency.org/honduras-how-a-surge-of- corruption-scandals-has-fueled-political-crisis-85af16ceac85?gi=93806c2b2d16 Armed Conflict Location & Event Data Project (ACLED). (2020). Central America and COVID-19: The Pandemic’s Impact on Gang Violence. https://acleddata.com/2020/05/29/central-america-and- covid-19-the-pandemics-impact-on-gang-violence/ ASJ. (2020). Coronavirus in Honduras . https://www.asj-us.org/learn/coronavirus-in-honduras BID. (2021). BIDeconomics Honduras: oportunidades de mayor desarrollo para todos en post pandemia. Inter-American Development Bank. https://doi.org/10.18235/0003885 Borgen Project. (2020). Political Corruption and Healthcare in Honduras - BORGEN. https://www.borgenmagazine.com/healthcare-in-honduras/ Brackett, R. (2020). Hurricane Eta Forces Red Alerts, Evacuations in Nicaragua and Honduras. The Weather Channel . https://weather.com/news/news/2020-11-02-hurricane-eta-nicaragua- honduras-evacuations-red-alerts Brekke, G. (2021). Chamelecón Revisited . https://www.maryknollmagazine.org/2021/06/chamelecon-revisited/ Brigada, A.-C. (2020). Coronavirus crisis exposes another pandemic in Honduras: Analysts | Coronavirus pandemic News | Al Jazeera. Al Jazeera. https://www.aljazeera.com/news/2020/3/27/coronavirus-crisis-exposes-another-pandemic-in- honduras-analysts 56 Cárdenas, C. (2012). Inventario y caracterización de los Sistemas de Alerta Temprana Honduras. https://www.eird.org/pr14/cd/documentos/espanol/AmericaCentralHerramientasydocumento s/AlertaTemprana/InventarioSATHonduras.pdf CARE, & UN Women. (2020). Latin America and the Carribaen Rapid Gender Analysis for COVID-19. https://reliefweb.int/report/brazil/latin-america-and-carribean-rapid-gender-analysis-covid-19 CARE, & UN Women. (2021). Rapid Gender Analysis in Honduras: an Overview in the Face of COVID- 19 and Eta/lota. https://reliefweb.int/report/honduras/rapid-gender-analysis-honduras- overview-face-covid-19-and-eta-lota Carmenate-Milián, L., Herrera-Ramos, A., Ramos-Cáceres, D., Lagos-Ordoñez, K., Lagos-Ordoñez, T., & Somoza-Valladares, C. (2017). Situation of the Health System in Honduras and the New Proposed Health Model. https://doi.org/10.21767/1989-5216.1000222 Cavallo, A., & Ireland, V. (2014). Preparing for Complex Interdependent Risks: A System of Systems Approach to Building Disaster Resilience. International Journal of Disaster Risk Reduction, 9, 181–193. https://doi.org/https://doi.org/10.1016/j.ijdrr.2014.05.001 CDC. (2021). Central American Refugee Health Profile. https://www.cdc.gov/immigrantrefugeehealth/profiles/central- american/index.html?CDC_AA_refVal=https%3A%2F%2Fwww.cdc.gov%2Fimmigrantrefugeehe alth%2Fprofiles%2Fcentral-american%2Fhealthcare-diet%2Findex.html#healthcare-access CENAOS. (n.d.). El Niño Oscilación del Sur (ENOS). Retrieved May 7, 2022, from http://cenaos.copeco.gob.hn/elnino.html CEPAL. (2021). Evaluación de los efectos e impactos de la tormenta tropical Eta y el huracán Iota en Honduras. cepal.org/es/publicaciones/46853-evaluacion-efectos-impactos-causados-la- tormenta-tropical-eta-huracan-iota CIA. (2022). Honduras - The World Factbook. The World Factbook. https://www.cia.gov/the-world- factbook/countries/honduras/ CIPPDV. (2015). Characterization of Internal Displacement in Honduras. Inter-Agency Commission for the Protection of Persons Displaced by Violence (CIPPDV). https://www.jips.org/uploads/2018/11/Honduras-ProfilingReport-InternalDisplacement-2015- EN.pdf Conde, M. (2021). Real time decisions during Hurricanes ETA and IOTA GEOGloWS-ECMWF, Honduras. AmeriGEO Week 2021. COPECO. (n.d.). Planes Municipales de Gestión de Riesgo - PGRD - Proyecto de Gestión de Riesgo de Desastres. Retrieved June 7, 2022, from https://www.pgrd-copeco.gob.hn/planes-municipales- de-gestion-de-riesgo/ Corson, B., & Hallock, J. (2020). Intersecting Crises: Pandemic and Hurricanes Add to Political Instability Driving Migration from Honduras. https://www.migrationpolicy.org/print/17101 Corson, B., & Hallock, J. (2021). Intersecting Crises: Pandemic and Hurricanes Add to Political Instability Driving Migration from Honduras FEATURE. Migration Policy. https://www.migrationpolicy.org/print/17101 Criterio. (2020, June 27). Más de un millón de hondureños sin recibir la “bolsa solidaria”, según datos del mismo gobierno. https://criterio.hn/mas-de-un-millon-de-hondurenos-sin-recibir-la-bolsa- solidaria-segun-datos-del-mismo-gobierno/ Dalby, C. (2020). Coronavirus Squeezing Finances of Mexico’s Criminal Groups. https://insightcrime.org/news/brief/coronavirus-squeezing-mexico-criminal-groups/ 57 de Ruiter, M. C., Couasnon, A., van den Homberg, M. J. C., Daniell, J. E., Gill, J. C., & Ward, P. J. (2020). Why We Can No Longer Ignore Consecutive Disasters. Earth’s Future, 8(3), e2019EF001425. https://doi.org/10.1029/2019EF001425 Dennis Álvarez. (2020). Primer frente frío dejaría temperaturas de 5 y 7 grados en algunas zonas de Honduras. HRN. https://www.radiohrn.hn/primer-frente-frio-dejaria-temperaturas-de-5-y-7- grados-en-algunas-zonas-de-honduras DG ECHO, & ACAPS. (2022). ECHO Daily Flash 13 May 2022. https://erccportal.jrc.ec.europa.eu/ECHO-Products/Echo-Flash#/daily-flash-archive/4492 Díaz-Bonilla, E., Flores, L., Paz, F., Piñeiro, V., & Zandstra, T. (2021). Honduras: The impact of covid-19 and policy implications: Second report. https://doi.org/10.2499/P15738COLL2.134533 DTM. (2021). Honduras-Baseline assessment of migration flows and mobility tracking within the context of COVID-19 report #6 (June, 2021) | DTM. https://dtm.iom.int/reports/honduras- baseline-assessment-migration-flows-and-mobiity-tracking-within-context-covid-19-3 el Diario. (2020). Más de 120 mil libras de alimentos están preposicionadas para atender a afectados por huracán Eta. El Diario. http://www.eldiario.hn/mas-de-120-mil-libras-de-alimentos-estan- preposicionadas-para-atender-a-afectados-por-huracan-eta/ Eligh, J. (2020). Crisis and Opportunity Impacts of the coronavirus pandemic on illicit drug markets. www.globalinitiative.net Elliott, S., Heinz, B., & Cashman, K. (2022). Honduras: Social and Economic Indicators After 12 Years of National Party Rule - Center for Economic and Policy Research. CEPR. https://cepr.net/honduras-social-and-economic-indicators-after-12-years-of-national-party- rule/ FAO. (2018). Dry corridor: FAO in Emergencies. https://www.fao.org/emergencies/crisis/dry- corridor/en/ FAO. (2019). Status of community-based forestry and forest tenure in Honduras. https://www.fao.org/3/ca5738en/CA5738EN.pdf FAO, & WFP. (2020). FAO y WFP preocupados por el grave impacto de la sequía entre los más vulnerables de Centroamérica | FAO. https://www.fao.org/americas/noticias/ver/es/c/1150346/ FEWS NET. (2021). El Salvador, Honduras, and Nicaragua - Key Message Update: Fri, 2021-01-29. https://fews.net/central-america-and-caribbean/el-salvador-honduras-and-nicaragua/key- message-update/january-2021 FEWS NET. (2011). Honduras Livelihood Profiles National Summary. FEWSNET. http://fews.net/sites/default/files/documents/reports/HN_Livelihoods%20Profile_EN.pdf FEWS NET. (2018). El Salvador, Honduras, and Nicaragua - Informe de monitoreo remoto . https://fews.net/central-america-and-caribbean/el-salvador-honduras-and-nicaragua/remote- monitoring-report/december FEWS NET. (2019). Central America and Caribbean: Key Message Update, August 2019 - Haiti | ReliefWeb. https://reliefweb.int/report/haiti/central-america-and-caribbean-key-message- update-august-2019 France 24. (2022). Cientos de nicaragüenses y hondureños parten en la primera caravana del año a EE.UU. 2022. https://www.france24.com/es/am%C3%A9rica-latina/20220115-caravana- migrante-honduras-nicaragua-eeuu Government of Honduras, & World Bank. (2019). Hydrometeorological and Climate Services Modernisation Plan for Honduras. In World Bank Group . 58 https://www.gfdrr.org/sites/default/files/publication/ExecutiveSummary_PMHidromet_Hondu ras_ENG.pdf GRC, Climate Centre, & IFRC. (2020). Forecast-based Financing Manual. https://manual.forecast- based-financing.org/ Hábitat para la Humanidad Honduras. (2019). Asentamientos irregulares en el Valle de Sula. www.lanoticia.hn Harrowsmith, M., Nielsen, M., Jaime, C., Coughlan de Perez, E., & Uprety, M. (2020). The Future of Forecasting: Impact based Forecasting for Early Action. https://www.forecast-based- financing.org/wp-content/uploads/2020/09/Impact-based-forecasting-guide-2020.pdf Herlihy, P. H., & Tappan, T. A. (2019). Recognizing Indigenous Miskitu Territory in Honduras. Geographical Review, 109(1), 67–86. https://doi.org/10.1111/GERE.12309 HRC. (2019). Estudio de Factibilidad sobre el uso de Forecast-Based Financing (FbF) en Honduras. HRC. (2021). Plan de Action Temprana - Inundaciones por Tormentas Tropicales. HRW. (2021). World Report 2021: Honduras . https://www.hrw.org/world-report/2021/country- chapters/honduras Human Rights Watch. (2019). World Report 2019: Events of 2018. https://reliefweb.int/attachments/2bfc6511-0b5a-3d1c-959f- 3281a6ba7924/hrw_world_report_2019.pdf Human Rights Watch. (2020). World Report 2020: Events of 2019. https://www.hrw.org/sites/default/files/world_report_download/hrw_world_report_2020_0. pdf Human Rights Watch. (2021). World Report 2021: Events of 2020. https://www.hrw.org/sites/default/files/media_2021/01/2021_hrw_world_report.pdf ICRC. (2017). Displaced in Cities : Experiencing and Responding to Urban Internal Displacement Outside Camps. https://shop.icrc.org/displaced-in-cities-experiencing-and-responding-to- urban-internal-displacement-outside-camps-pdf-en.html IDMC. (2016). A web of violence. Crime, corruption and displacement in Honduras. https://www.internal-displacement.org/sites/default/files/publications/documents/201903- honduras-web-of-violence-en.pdf IDMC. (2020). Honduras: Displacement associated with Conflict and Violence. International Displacement Monitoring Centre. https://www.internal-displacement.org/countries/honduras IFRC. (2019). Dengue spreads across Central America, Red Cross scales up response - Honduras. https://reliefweb.int/report/honduras/dengue-spreads-across-central-america-red-cross- scales-response IFRC. (2020a). Central America: Dengue Outbreak Emergency Plan of Action, 6-months Operations update (MDR42005). https://reliefweb.int/report/honduras/central-america-dengue-outbreak- emergency-plan-action-6-months-operations-update IFRC. (2020b). Honduras: Hurricane Eta and Iota - Emergency appeal n° MDR43007 Operation Update no. 2 - Honduras. https://reliefweb.int/report/honduras/honduras-hurricane-eta-and- iota-emergency-appeal-n-mdr43007-operation-update-no-2 IFRC. (2021). Communities affected by Hurricanes Eta and Iota are threatened by food insecurity, displacement and the climate crisis. https://www.ifrc.org/press-release/communities-affected- hurricanes-eta-and-iota-are-threatened-food-insecurity INFORM. (2021). Inform Honduras. https://drmkc.jrc.ec.europa.eu/inform-index/INFORM- Subnational-Risk/Honduras 59 International Crisis Group. (2020). Virus-proof Violence: Crime and COVID-19 in Mexico and the Northern Triangle. https://www.crisisgroup.org/latin-america-caribbean/83-virus-proof- violence-crime-and-covid-19-mexico-and-northern-triangle IPC. (2022). Honduras: Integrated Food Security Phase Classification Snapshot, December 2021 - August 2022. https://reliefweb.int/report/honduras/honduras-integrated-food-security-phase- classification-snapshot-december-2021-august IPCC. (2021). Regional fact sheet – North and Central America. https://www.ipcc.ch/report/ar6/wg1/downloads/factsheets/IPCC_AR6_WGI_Regional_Fact_Sh eet_North_and_Central_America.pdf Olson, J. (2020). Honduran storm survivors form US-bound migrant caravan. The New Humanitarian. https://www.thenewhumanitarian.org/news-feature/2020/12/11/honduras-hurricane- survivors-migrant-caravan Kopp, S., & Meriam, E. (2020). Significant Improvements to Global Streamflow Services. ESRI ArcGIS. https://www.esri.com/arcgis-blog/products/arcgis-living-atlas/water/global-stream-flow-2020/ Kossin, J. P., Hall, T., Knutson, T., Kunkel, K. E., Trapp, R. J., Waliser, D. E., & Wehner, M. F. (2017). Ch. 9: Extreme Storms. Climate Science Special Report: Fourth National Climate Assessment, Volume I. https://doi.org/10.7930/J07S7KXX Latin American and Caribbean Network of Fair Trade Small Producers and Workers (CLAC). (2020). Report on the impact of hurricanes ETA and IOTA in Central America . https://clac- comerciojusto.org/en/2020/11/informe-sobre-impacto-de-los-huracanes-eta-e-iota-por- centroamerica/ LAWG. (2020). How Corruption Fuels Human Rights Abuses & Forced Migration: The Case of Honduras - Latin America Working Group. https://www.lawg.org/corruptionhonduras/ Marchezini, V., & Wisner, B. (2017). Challenges for vulnerability reduction in Brazil: Insights from the PAR framework. In Reduction of vulnerability to disasters: from knowledge to action. São Carlos: Rima Editora. https://www.academia.edu/34884801/Challenges_for_vulnerability_reduction_in_Brazil_Insig hts_from_the_PAR_framework_In_Reduction_of_vulnerability_to_disasters_from_knowledge _to_action_edited_by_Victor_Marchezini_Ben_Wisner_Luciana_de_Resende_Londe_and_Silvi a_Midori_Saito_p_53_92_S%C3%A3o_Carlos_Rima_Editora_2017 Marin-Ferrer, M., Vernaccini, L., & Poljansek, K. (2017). INFORM Index for Risk Management Concept and Methodology . https://doi.org/10.2760/094023 Martheswaran, T. K., Hamdi, H., Al-Barty, A., Zaid, A. A., & Das, B. (2022). Prediction of dengue fever outbreaks using climate variability and Markov chain Monte Carlo techniques in a stochastic susceptible-infected-removed model. Scientific Reports 2022 12:1, 12(1), 1–17. https://doi.org/10.1038/s41598-022-09489-y Masters, J. (2019). Fifth Straight Year of Central American Drought Helping Drive Migration. Scientific American. https://blogs.scientificamerican.com/eye-of-the-storm/fifth-straight-year-of-central- american-drought-helping-drive-migration/ McSweeney, K., Wrathall, D. J., Nielsen, E. A., & Pearson, Z. (2018). Grounding traffic: The cocaine commodity chain and land grabbing in eastern Honduras. Geoforum, 95, 122–132. https://doi.org/10.1016/J.GEOFORUM.2018.07.008 Moloney, A. (2019). In Honduras, years of drought pressure farmers to leave land. https://www.reuters.com/article/us-climate-change-honduras-migrants-anal- idUSKBN1WC0M6 60 MSF. (2019). MSF expands efforts to control dengue epidemic, Honduras. https://www.msf.org/msf- expands-efforts-control-dengue-epidemic-honduras National Hurricane Center, & Central Pacific Hurricane Center. (2020). Hurricane ETA. In NOAA. https://www.nhc.noaa.gov/archive/2020/al29/al292020.public_a.007.shtml OCHA. (2019). Latin America & the Caribbean: Weekly Situation Update, as of 29 July 2019. https://reliefweb.int/report/world/latin-america-caribbean-weekly-situation-update-22-29- july-2019-29-july-2019 OCHA. (2020a). Central America: 2020 Hurricane Season - Situation Report No. 3 (as of 8:00am EST 16 November 2020). United Nations . www.unocha.org OCHA. (2020b). Situation Overview Hurricane Eta and Iota - Honduras, Nov 2020. https://vosocc.unocha.org/GetFile.aspx?xml=7034SlrCTans7AFD3Old6HY60h65GGfkhSDSLr02P daP2Ewx_l1.html&tid=7034&laid=1 OCHA. (2022). Hurricanes Eta e Iota: seguimento nacional de albergues . https://app.powerbi.com/view?r=eyJrIjoiOGYwYTE2NWEtYTg3Ni00ZjIwLTliOWUtY2U0OWVjNz YyOTU0IiwidCI6IjBmOWUzNWRiLTU0NGYtNGY2MC1iZGNjLTVlYTQxNmU2ZGM3MCIsImMiOjh9 Oliver-Smith, A., Alcántara-Ayala, I., Burton, I., & Lavell, A. (2016). A conceptual framework and guide to research Forensic Investigations of Disasters (FORIN) Forensic Investigations of Disasters (FORIN) A conceptual framework and guide to research. www.irdrinternational.org Orellana, D., & Arita, D. (2021, October 8). Eta e IOTA hundieron a Chamelecón, es urgente ayudarles . Diario En Alta Voz. https://enaltavoz.com/eta-e-iota-hundieron-a-chamelecon-es-urgente- ayudarles/ PAHO. (2019). PAHO/WHO Data - Incidence. WHO. https://www3.paho.org/data/index.php/en/mnu-topics/indicadores-dengue-en/dengue- regional-en/315-reg-dengue-incidence-en.html?start=1 PAHO - WHO. (2019). PAHO/WHO Data - Honduras - Dengue Cases. PAHO/WHO. https://www3.paho.org/data/index.php/en/mnu-topics/indicadores-dengue-en/dengue- subnacional-en/543-hnd-dengue-casos-en.htmlpahadataondenguegoesuntilmid-2019 Palencia, G. (2021). Honduras’ Castro to seek new debt deal with IMF, opposes tax hikes | Reuters. Reuters. https://www.reuters.com/markets/rates-bonds/honduras-castro-eyes-new-debt-deal- with-imf-opposes-new-taxes-2021-11-30/ Pascale, S., Kapnick, S. B., Delworth, T. L., Hidalgo, H. G., & Cooke, W. F. (2021). Natural variability vs forced signal in the 2015–2019 Central American drought. Climatic Change, 168(3–4), 1–21. https://doi.org/10.1007/S10584-021-03228-4/FIGURES/9 Pasch, R. J., Reinhart, B. J., Berg, R., & Roberts, D. P. (2021). National Hurricane Center Tropical Cyclone Report Hurricane Eta . https://www.nhc.noaa.gov/data/tcr/AL292020_Eta.pdf Patricola, C. M., & Wehner, M. F. (2018). Anthropogenic influences on major tropical cyclone events. Nature 2018 563:7731, 563(7731), 339–346. https://doi.org/10.1038/s41586-018-0673-2 Pescaroli, G., & Alexander, D. (2016). Critical infrastructure, panarchies and the vulnerability paths of cascading disasters. Natural Hazards, 82(1), 175–192. https://doi.org/10.1007/S11069-016- 2186-3/FIGURES/3 Potter, C. (2019). Honduras Fights “Worst Dengue Outbreak in Decades”, Outbreak Observatory. https://www.outbreakobservatory.org/outbreakthursday-1/8/1/2019/honduras-fights-worst- dengue-outbreak-in-decades Project HOPE. (2020). Hurricanes Eta & Iota. https://www.projecthope.org/hurricanes-eta-iota- contaminated-water-honduras/12/2020/ 61 Pulte Institute. (2020). Informe de Migración de Hondureños en el Contexto de la Pandemia Covid-19. https://pulte.nd.edu/assets/424803/ciprodeh_migracion.pdf Recio, M. E. (2015). Honduras Resource Tenure and Sustainable Landscapes Assessment. USAID Tenure and Global Climate Change Program. https://www.land-links.org/wp- content/uploads/2016/09/USAID_Land_Tenure_TGCC_Honduras_Assessment_En.pdf Redacción Web. (2020). Policía se suma a emergencias provocadas por Eta. Primicia Honduras . https://primiciahonduras.hn/policia-se-suma-a-emergencias-provocadas-por-eta/ ReliefWeb. (2019). Central America and the Caribbean: Drought - Jul 2018. https://reliefweb.int/disaster/dr-2018-000106-slv Reuters. (2019). Former first lady of Honduras sentenced to 58 years in jail. https://www.reuters.com/article/us-honduras-corruption-idUSKCN1VQ00Y Reuters. (2021). La crisis del café en Centroamérica impulsa migración récord hacia el norte. Euronews. https://es.euronews.com/2021/12/08/migracion-cafe-centroamerica Semple, K. T. (2019). Climate Change and Political Chaos: A Deadly Mix in Honduras Dengue Epidemic. The New York Times. https://www.nytimes.com/2019/12/29/world/americas/honduras-dengue-epidemic.html SEPOL. (2022). Violence data Honduras. https://www.sepol.hn/ Spring, K., & Ceja, J. L. G. (2020). Episode 13: Hurricanes Iota and Eta: Not Just Natural Disasters. Honduras Now. https://www.hondurasnow.org/episode-13-hurricanes-iota-and-eta-not-just- natural-disasters/ SPS. (2017). Plan de Emergencia Municipal. https://www.pgrd-copeco.gob.hn/wp- content/uploads/2019/07/PEM-SPS.pdf Stewart, I. T., Maurer, E. P., Stahl, K., & Joseph, K. (2022). Recent evidence for warmer and drier growing seasons in climate sensitive regions of Central America from multiple global datasets. International Journal of Climatology, 42(3), 1399–1417. https://doi.org/10.1002/JOC.7310 Stewart, S. R. (2021). National Hurricane Center Tropical Cyclone Report Hurricane Iota. https://www.nhc.noaa.gov/data/tcr/AL312020_Iota.pdf Telam. (2022). Cientos de familias migrantes partieron desde Honduras en una nueva caravana hacia EEUU. National Hurricane Center Tropical Cyclone Report Hurricane Iota Tellman, B., Sesnie, S. E., Magliocca, N. R., Nielsen, E. A., Devine, J. A., McSweeney, K., Jain, M., Wrathall, D., Benessaiah, K., Aguilar-Gonzalez, B., & Author, C. (2020). Illicit Drivers of Land Use Change: Narcotrafficking and Forest Loss in Central America. https://www.sciencedirect.com/science/article/pii/S0959378019313287 Thompson, A., & Montañez, A. (2020). In 2020, Record-Breaking Hurricanes Arrived Early and Often. Scientific American. https://www.scientificamerican.com/article/in-2020-record-breaking- hurricanes-arrived-early-and-often1/ Tiempo. (2020). Sula Valley begins to reactivate after eta and Iota. https://tiempo.hn/valle-de-sula- comienza-a-reactivarse-tras-paso-de-eta-e-iota/ UNAH - IUDPAS. (2019). Boletines Nacionales de la Violencia. Universidad Nacional Autónoma de Honduras (UNAH); Instituto Universitario En Democracia, Paz y Seguridad (IUDPAS). https://iudpas.unah.edu.hn/observatorio-de-la-violencia/boletines-del-observatorio- 2/boletines-nacionales/ UNDP. (2021). Análisis sobre la situación de la violencia y seguridad ciudadana en Honduras durante 2020. https://www.hn.undp.org/content/honduras/es/home/presscenter/articles/2021/analisis- multidimensional-de-la-seguridad-ciudadana-en-honduras-.html 62 UNDRR. (2022). Scoping Study On Compound, Cascading And Systemic Risks In The Asia Pacific. UNICEF. (2020). Over 1.2 million children affected by ETA across Central America. https://www.unicef.org/press-releases/over-12-million-children-affected-eta-across-central- america-unicef UNSDG. (2021). Violence against women, the other pandemic impacting Honduras. https://unsdg.un.org/latest/stories/violence-against-women-other-pandemic-impacting- honduras US Department of Justice. (2018). Former Honduran Congressman And Brother Of The Current President Of Honduras Charged With Conspiring To Import Cocaine Into The United States And Related Firearms Offenses. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved =2ahUKEwiP4- HorK_5AhUQ9aQKHWyNBzcQFnoECAUQAQ&url=https%3A%2F%2Fwww.justice.gov%2Fusao- sdny%2Fpr%2Fformer-honduran-congressman-and-brother-current-president-honduras- charged-conspiring&usg=AOvVaw26ThQ5-2KDzDS4-HFD9r0D US Embassy Tegucigalpa. (2022). U.S. Actions against Former Honduran President Juan Orlando Hernandez for Corruption - U.S. Embassy in Honduras. https://hn.usembassy.gov/u-s-actions- against-former-honduran-president-juan-orlando-hernandez-for-corruption/ USAID. (2017). Climate Change Risk Profile Honduras. Valladares, H. R. M. (2022). FAH lleva alimentos a La Mosquita a damnificados por el huracán Eta. La Noticia HN. https://lanoticia.hn/fah-lleva-alimentos-a-la-mosquita-a-damnificados-por-el- huracan-eta/ Vazquez, G. (2020). Contagion and extortion in Central America | Global Initiative. Global Initiative against Transnational Organized Crime . https://globalinitiative.net/analysis/covid-extortion- central-america/ Verza, M. (2021). Desperation grows in battered Honduras, fueling migration. AP News. https://apnews.com/article/honduras-hurricane-iota-mexico-storms-immigration- 3cf340e556ee767d1dd3dce351c934b5 WFP. (2017). Food Security and Emigration: Why people flee and the impact on family members left behind in El Salvador, Guatemala and Honduras. https://docs.wfp.org/api/documents/WFP- 0000019632/download/ WFP. (2018). El Niño response in the Dry Corridor of Central America: an evaluation | World Food Programme. https://www.wfp.org/publications/el-ni%C3%B1o-response-dry-corridor-central- america-evaluation WFP. (2022). Honduras Country Brief. www.wfp.org/countries/honduras Wisner, B., Blaikie, P., Cannon, T., & Davis, I. (2014). At risk: natural hazards, peoples vulnerability and disasters. At Risk: Natural Hazards Peoples Vulnerability and Disasters, 1–471. https://doi.org/10.4324/9780203714775 World Bank. (2017). Setting a Historical Precedent for Land Rights in Honduras: Improving Decentralized Land Administration Services. Results Briefs. https://www.worldbank.org/en/results/2017/10/30/setting-a-historical-precedent-for-land- rights-in-honduras-improving-decentralized-land-administration-services World Bank. (2020). Project Information Document - Honduras Tropical Cyclones Eta and Iota Emergency Recovery Project (P175977). https://view.officeapps.live.com/op/view.aspx?src=https%3A%2F%2Fdocuments1.worldbank.o rg%2Fcurated%2Fen%2F656441607382006563%2FProject-Information-Document-Honduras- 63 Tropical-Cyclones-Eta-and-Iota-Emergency-Recovery-Project- P175977.docx&wdOrigin=BROWSELINK World Bank. (2021a). Honduras Macro Poverty Outlook. https://thedocs.worldbank.org/en/doc/e408a7e21ba62d843bdd90dc37e61b57- 0500032021/related/mpo-hnd.pdf World Bank. (2021b). Poverty & Equity Brief Honduras. www.worldbank.org/poverty World Bank. (2022a). Mobile cellular subscriptions (per 100 people) - Honduras. https://data.worldbank.org/indicator/IT.CEL.SETS.P2?locations=HN World Bank. (2022b). Population, total - Honduras. https://data.worldbank.org/indicator/SP.POP.TOTL?locations=HN World Bank. (2022c, April 25). Honduras: panorama general. https://www.worldbank.org/en/country/honduras/overview World Bank. (2022d, April 25). Honduras: panorama general. World Bank Group. (2018). Global Crisis Risk Platform. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved =2ahUKEwjH8IXuq6_5AhX0h_0HHWCAAHcQFnoECAsQAw&url=https%3A%2F%2Fdocuments.w orldbank.org%2Fen%2Fpublication%2Fdocuments- reports%2Fdocumentdetail%2F660951532987362050%2Fglobal-crisis-risk- platform&usg=AOvVaw1Rgg6DXqwGH-YKxjem4ciy World Meteorological Organization. (2018). Multi-hazard Early Warning Systems: A Checklist, Outcome of the first Multi-hazard Early Warning Conference. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved =2ahUKEwiA- uzSq6_5AhUNgf0HHQXMCBIQFnoECAsQAQ&url=https%3A%2F%2Flibrary.wmo.int%2Fdoc_nu m.php%3Fexplnum_id%3D4463&usg=AOvVaw0zYYkEj1E2LmLO4Q3N3ZaH Zambrano, L. I., Fuentes-Barahona, I. C., Henriquez-Marquez, K. I., Vasquez-Bonilla, W. O., Sierra, M., Muñoz-Lara, F., Luna, C., Bonilla-Aldana, D. K., & Rodriguez-Morales, A. J. (2021). COVID-19 and Hurricanes: The Impact of Natural Disasters during a Pandemic in Honduras, Central America. In Prehospital and Disaster Medicine (Vol. 36, Issue 2, pp. 246–248). Cambridge University Press. https://doi.org/10.1017/S1049023X21000182 Zambrano, L. I., Rodriguez, E., Espinoza-Salvado, I. A., Fuentes-Barahona, I. C., Lyra de Oliveira, T., Luciano da Veiga, G., Cláudio da Silva, J., Valle-Reconco, J. A., & Rodríguez-Morales, A. J. (2019). Spatial distribution of dengue in Honduras during 2016–2019 using a geographic information systems (GIS)–Dengue epidemic implications for public health and travel medicine. Travel Medicine and Infectious Disease, 32, 101517. https://doi.org/10.1016/J.TMAID.2019.101517 Zscheischler, J., Martius, O., Westra, S., Bevacqua, E., Raymond, C., Horton, R. M., van den Hurk, B., AghaKouchak, A., Jézéquel, A., Mahecha, M. D., Maraun, D., Ramos, A. M., Ridder, N. N., Thiery, W., & Vignotto, E. (2020). A typology of compound weather and climate events. Nature Reviews Earth & Environment 2020 1:7, 1(7), 333–347. https://doi.org/10.1038/s43017-020- 0060-z Annexes 64 Annex 1: Overview of hurricanes Eta and Iota genesis, track, strength and direct impacts Eta was classified as a hurricane on 2 November 2020. It strengthened rapidly to Category 4 intensity on the same day. Late on 3 November, Eta made landfall off the eastern coast of Nicaragua with a storm surge of 26–33 feet above normal sea levels (Pasch et al., 2021). The hurricane moved west across Nicaragua on 4 November, weakening to a tropical storm and then a tropical depression by 5 November, when it crossed the border into Honduras (A1 Figure 1). Due to the slow motion of the cyclone, it dumped heavy amounts of rainfall across the country and especially in the northern regions of Honduras (A1 Figure 2). From 1–7 November, rainfall totals of 31.63 inches were reported in Tela and 29.25 inches in La Ceiba (Pasch et al., 2021). This extreme rainfall caused flash flooding, river flooding and landslides across Central America, and contributed to at least 74 deaths in Honduras (Pasch et al., 2021). The storm moved north-westwards across Honduras on 5 November, before strengthening slightly as it emerged into the Gulf of Honduras and continued northwards. A1. Figure 1 Hurricane Eta storm track and windspeeds. Source: Map by the authors, data: EC-JRC; NOAA (2020). 65 A1. Figure 2. Hurricane Eta rainfall amounts. Source: IDB, 2021. Meanwhile, a second storm was brewing in the Caribbean. Hurricane Iota formed in a similar way to Hurricane Eta around 12 November and strengthened into a hurricane over the next four days. It was assessed by the National Hurricane Center as a strong Category 4 hurricane by 15 November (S. R. Stewart, 2021). This marked the first time two major hurricanes had been recorded in November, which is typically the end of the hurricane season. On 17 November, Iota made landfall on the eastern coast of Nicaragua. It then made its way west and north, losing speed and power over the north-west portion of Nicaragua as it crossed the border into Honduras. The hurricane dropped heavy rains on a landscape already saturated by the flooding from Eta. In Honduras, La Ceiba received 11.85 inches while Santa Bárbara, in the west of the country, received 8.90 inches (Pasch et al., 2021). Continuing to move west, Iota dissipated over El Salvador by late in the day on 18 November (Figure 4). Thirteen direct deaths were attributed to the hurricane, mostly due to mudslides. In Copán Ruinas, a town in the far west of the country, 80 per cent of the roads were impassable, and the Ramón Villeda Morales International Airport complex was submerged for a month under 1 metre of floodwater (S. R. Stewart, 2021). 66 A1. Figure 3. Hurricane Iota rainfall amounts. Source: IDB, 2021. 67 A1. Figure 4 . Hurricane Iota track and windspeeds. Source: Map by the authors, data: EC-JRC; NOAA (2020) The two maps below show the extent of satellite-detected surface waters (cumulative) in Honduras in 13–17 November 2020 (after Hurricane Eta made landfall) and 25–29 November 2020 (after Hurricane Iota made landfall). The maps illustrate that the heavy rainfalls that were caused by the two hurricanes caused severe flooding in the north-west, east and south of Honduras. It shows that between the two time periods the flooding intensified especially in the north-east and north of the country as well as around the Sula Valley. Most of the flooded areas were primarily not within the direct impact-zones of the hurricanes, highlighting the complexity in space and time of hurricane impacts. A1. Figure 5. Flood extent from 13-17 November across Honduras. Map by authors, data: UNOSAT/NOAA (2020) 68 A1. Figure 6. Flood extent 25-29 November 2020 in Honduras. Map by authors, data: UNOSAT/NOAA (2020) A1. Figure 7. Flood extent November 2020 in the Sula Valley, Honduras. Map by authors, data: UNOSAT/NOAA (2020) Annex 2: Hurricanes Eta and Iota forecast analysis The following section explores to what extent Hurricanes Eta and Iota were forecast accurately and in a timely manner, looking at track, wind speed and rainfall. The genesis of the timing and location of Hurricane Eta was not that well predicted compared with Hurricane Iota, which was forecast exceptionally well. The challenge in forecasting Eta was the erratic track of the storm and the rapid 69 intensification as it approached land (See A2 Figures 1a and b). A verification of National Hurricane Center–NOAA official track forecasts for Hurricane Eta indicated errors were a little higher than the mean official errors for the previous five-year period – a reasonably good performance considering the erratic nature of the tropical cyclone’s track. Several of the guidance models had lower mean errors than the official forecasts, but in most of those cases the NHC forecasts had comparable errors. However, some models had substantially lower errors than most of the other track guidance at 48 hours (h) and beyond. Nonetheless, all models in the official forecasts consistently predicted landfall in north-east Nicaragua. The change in track of Eta meant that the north and north-west was affected by the storm for longer, compared to initial forecasts (e.g., forecast issue #7 in Figure 1a). A2. Figure 1a. Hurricane Eta, forecast #7 issued 2 November 2020. Source: Authors, using Tropycal Python tool 70 A2. Figure 1b. Hurricane Eta, forecast #15 issued 4 November 2020. Source: Authors, using Tropycal Python tool) The verification of NHC official wind intensity forecasts for Eta, shows that these values were somewhat higher than the mean official errors for the previous five-year period for the 12–48h forecast intervals, and lower than the five-year means at 60–120h. In general, however, the mean official errors were comparable to, or lower than, those of the intensity guidance models. The rapid intensification of Eta before landfall in Nicaragua was problematic for forecasting models. The official intensity forecasts issued 30h before the Nicaragua landfall showed significant strengthening and indicated that the system would become a major hurricane by the time it reached the coast (Fig. 4a). “[About Hurricane Eta] If now if you ask me as a meteorology professional, it was certain that I was going to enter, there was no doubt about that, the doubt was regarding the territory in the trajectory of the hurricane.” Academic 1 Hurricane Eta not only brought strong winds, but it also resulted in very heavy precipitation in the north and north-west of Honduras, triggering flooding and landslides especially in the Cortés, Copán and Yoro departments. Heavy rainfall was forecast very well by many agencies, including CENAOS, while ECMWF had good performance seven days in advance of the rainfall peak. The rainfall forecasts did indicate that heavy rainfall would concentrate in the north-west of Honduras (Figure 3). NOAA’s Outlook mentioned on 2 November: “Eta is expected to produce the following rainfall amounts through Friday evening: much of Nicaragua and Honduras: 15 to 25 inches (380 to 635 mm)”. These amounts turned out to be much higher in the coastal areas of Atlántida – on the northern coast of Honduras 31.63 inches were recorded (803.3 mm) in Tela and 29.25 inches (743.0 mm) at Golosón International Airport in La Ceiba from 1–7 November 2020 (Pasch et al., 2021). 71 A2. Figure 2. ECMWF heavy rainfall forecast. Shows probability (%) of exceeding 300 millimetres (mm) of accumulated rainfall over the forecast range of 10 days for the ensemble ECMWF forecast. Source: ECMWF image by GloFAS. As mentioned, Iota was forecast very well. A verification of NHC official track forecasts for Iota showed the official forecast track errors (A2 Figure 3a & A2. Figure 3b) were lower than the mean official errors for the previous five-year period at all forecast times. Although the official forecast track errors were better than average, so were the climatology track errors, which is an indication that Iota was relatively easy to forecast. The forecasts were clear on the eventual landfall location along the east- central coast of Nicaragua up to 54 hours before. 72 A2. Figure 3a. Hurricane Iota, forecast #7 issued 15 November 2020. Source: Authors, using Tropycal Python tool A2. Figure 3b. Hurricane Iota, forecast #15 issued 17 November 2020. Source: Authors, using Tropycal Python tool The verification of NHC official intensity forecasts for Iota, showed forecast intensity errors were lower than the mean official errors for the previous five-year period at the 24, 36, 72, 96 and 120h forecast periods. NHC forecasters performed well in predicting the first 30h, and the forecasts also correctly captured the rapid weakening trend at and after Iota’s landfall. The official wind intensity errors with some models gave intensity errors generally lower than all the available intensity guidance (A3 Figure 73 3b). Overall, the track and intensity of Iota were well forecast. It should also be noted that a major challenge for national-level forecasting of windspeed and rainfall was the lack of field measurement stations, as Eta had already damaged a lot of equipment (Humanitarian worker 4, Academic 2 and S. R. Stewart, 2021) A2. Figure 4a. Eta forecast intensity error (kt): Source: Authors, data from NHC-NOAA A2. Figure 4b. Hurricane Iota, forecast intensity error (kt): Produced by authors, data from NHC-NOAA The wind watches and warnings associated with Eta were issued for the north-east coast of Nicaragua about 36h before the onset of tropical storm force winds over that area, and 54h prior to landfall. This indicates a relatively short lead time; although it should be noted that Eta only reached Honduras 74 later. For Iota, Tropical Storm Watches and Warnings were issued well in advance. The warnings were verified with lead times of at least 51h and 42h, respectively, while the Hurricane Watch and a Hurricane Warning verified with lead times of approximately 54h and 45h, respectively. (Kopp & Meriam, 2020)(Kopp & Meriam, 2020)(Kopp & Meriam, 2020)It should be noted that there were no forecasts available of the consequences of Eta and Iota – that is, for the losses and damages resulting from the strong winds, heavy rainfall, riverine flooding and landslides. Furthermore, there were no national or international flood alerts shared, although the international GloFas model was publicly available for the relevant rivers (Kopp & Meriam, 2020). Early warning systems in Honduras for floods and landslides are typically organized at the community scale (Cárdenas, 2012). For floods, this involves monitoring the amount of rain over a particular period, or the river levels based on painted scales on bridges, carried out by volunteers who alert neighbours and the municipality (Civil society organization 1). Several KIs mentioned that the flood onset was a lot more rapid and intense in the Ulúa and Chamelecón basins, defying the “common sense” of those monitoring the river levels and catching communities by surprise (Humanitarian worker 6). Annex 3: Overview of disaster impacts Eta and Iota Sector Subsector Economic damage (USD million) Social $ 189,240 Education $ 19,560 Health $ 30,072 Housing $ 139,608 Productive $ 237,408 Agriculture $ 42,048 Tourism $ 11,352 Commerce & industry $ 184,008 Infrastructure $ 79,944 Electricity $ 2,616 Water & sanitation $ 27,888 Transport $ 49,176 Telecommunications $ 264 Environment $ 37,608 Total $ 544,200 A3. Summary Table 1: Economic damages by sector, source: CEPAL, 2021. Disaster Details Areas affected Impact Mortality 95 people lost their lives as a result of Cortés, Santa Bárbara and drowning and burying/taken away by mud- Lempira districts and landslides: 10 people missing (IDB, 2021) 75 Loss and 90,000 homes damaged by floods and Atlantic coast, Valle de Sula, damage of mud/landslides; 170.000 people evacuated Valle de Lean and Valle del homes and (CEPAL, 2021). 88,000 people have been Aguán – especially around San other identified as in need of temporary shelter and Pedro Sula, along rivers household 35,000 people needing house repairs (OCHA, Chamelecón and Ulúa assets 2020b) Areas isolated A total of 927 roads were affected and more Colón district up to several than 72 bridges damaged, while 62 bridges weeks were destroyed. As roads and bridges were washed away, 368,901 people were isolated for several weeks (IFRC, 2020a) Reduced 2 million people with limited or no access to Cortés (70% of damaged health access to health services and at increased risk of facilities and 77% of people with healthcare contracting COVID-19. More than a month limited access to healthcare) after the storms hit, several thousand people remained without access to clean water and health services (Project HOPE, 2020). As of 1 Dec 2020, 400 health facilities had reported damage, of which 120 were reported inoperative. At least 12 health facilities reported damage to their cold chain equipment, which has disrupted the refrigeration of critical medicines and vaccines (Project HOPE, 2020) Flooding and mudslides damaged health infrastructure, health workers were affected, transport networks disturbed isolating communities Loss of 3% of schools in Honduras damaged, 3% used Damage to educational centres schools as shelter (IDB, 2021) – only a fraction were occurred mostly in Cortés, permanently destroyed. Combination of Atlántida and in Francisco damage to school buildings and use of schools Morazán (IDB, 2021). The use of for shelter. Reports indicate increased drop- schools for shelter was out rates concentrated in Sula Valley and El Paraíso district (IDB, 2021). Permanent destruction in Atlántida and Cortés Limited access Almost all the sanitation infrastructure in the to WASH communities along the Chamelecón and Ulúa facilities rivers was destroyed, and many community wells were damaged as a result of flooding and mudslides (IFRC, 2021a) 76 Increase in Stagnant water, damage to WASH Especially in Sula Valley (along communicable infrastructure, poor shelter conditions, lack of Chamelecón and Ulúa rivers) diseases access to safe drinking water and sanitation (especially increased transmission and exposure to diarrhoea) waterborne diseases and dengue As of 1 Dec, several thousand Hondurans remain without access to clean water and health services. Most water sources in Santa Bárbara are contaminated, exposing the population to illnesses such as waterborne and skin diseases Loss of Losses of up to 80% in the agricultural sector Sula Valley, although impacts agricultural due to damages from both tropical storms across the country. production (IFRC, 2020a). The agricultural crops with the greatest damage in 2020 were: coffee (49%), Northern Atlantic departments banana (27%), plantain (7%) and sugarcane of Atlántida, Cortés, Santa (7%) Bárbara and Yoro. Southern and western parts of the country Further losses to banana (7%) and sugar cane (Choluteca, Francisco Morazán, (5%) were incurred through a reduction in the El Paraíso, Santa Bárbara, planted area, loss of equipment and therefore Lempira, Copán and a reduction in productivity (IDB, 2021). Ocotepeque). Also, Gracias a Dios in the north-east. 2.9 million people faced acute food insecurity at crisis or worse levels (Integrated Food Security Phase Classification [IPC], December 2020–March 2021; FAO & WFP, 2020). GBV and other Populations in need of greater protection Sula Valley shelters and areas protection included women, girls, adolescents, people impacted by the hurricanes concerns with disabilities, LGBTQI+ people and coincide with those inhabited by indigenous and Afro-descendent populations. the latter, including Afro- Women and girls were exposed to sexual Honduran Garifuna, Tawahka, abuse in shelters and gender-based violence and Miskito IPs (IFRC, 2021a). (GBV) rates have increased (CARE & UN Women, 2020). A3. Summary Table 2. Combined humanitarian impacts of hurricanes Eta and Iota in Honduras (sources: IDB, 2021; IFRC, 2020a) 77 Annex 4: Pressure and Release Frameworks for main impacts of Eta/Iota Root cause analysis disaster Impact: Shelter sector A4. Figure 1. Pressure and Release diagram (modelled on Wisner et al. 2014) for shelter-related primary, secondary and tertiary impacts. Please note that lack of coping capacity is not included in this visualization but should be considered in the future. 78 A4. Figure 1a. Pressure and Release diagram for health impacts A4. Figure 1b. Pressure and Release diagram for food insecurity. 79 Annex 5: Overview of available data sources for anticipation of crises in Honduras For the various risks factors (hazards, vulnerability and exposure), dynamic pressures and root causes identified in this study that contributed to the disaster impacts observed, the following (inter)national information sources could have been utilized: ▪ Forecasts of the hurricanes were available through the National Hurricane Center (see Annex 1and 2). This is monitored nationally by CENAOS. ▪ (Kopp & Meriam, 2020)(Kopp & Meriam, 2020)(Kopp & Meriam, 2020)For flooding, information on peak flows was available through the GloFas model (Kopp & Meriam, 2020)and Empresa Nacional de Energía Eléctrica (ENEE) (National Electricity Company) had information on reservoir levels, predicted peaks and projections for reservoir capacity (Conde, 2021). Furthermore, there were general warnings of heavy rainfall in the Sula Valley and awareness of general areas at risk. However, the breaching of some flood barriers was an unexpected event that was not anticipated. This is monitored by CENAOS. ▪ Landslides: no forecasting tools available, only an understanding of most at risk based on indicators such as slope, drainage infrastructure, soil type. ▪ Health risks, especially communicable diseases were monitored by the Ministry of Health, the Pan American Health Organization (PAHO) and international organizations, but no predictive tools were in use at the time. • Dengue: There is and was no epidemiological early warning system in place for dengue, although experts in retrospect have carried out predictive analysis on the Honduras crisis, and found that previous chikungunya, zika and dengue occurrence, combined with favourable climatic conditions, served as the best predictors (Martheswaran et al., 2022; Zambrano et al., 2019). In the 2019 case, the dry and warm conditions of 2018 have been indicated as an early warning sign for the potential for increased activity of dengue vectors. PAHO published information on areas most affected by dengue, showing high risk in Cortés(PAHO - WHO, 2019). MSF reports that, although alerts are there, the 2019 outbreak reached crisis levels because of high insecticide-resistance, lack of response, lack of health system capacity and low awareness of preventive measures, which limits early warning effectiveness (MSF, 2019). • COVID-19: Despite the prevalence of COVID-19 in other countries, there were no early warning signals in Honduras. The situation was monitored on a daily and weekly basis by the Ministry of Health, although several sources question the accuracy of the statistics, given the lack of testing and reporting infrastructure in Honduras – especially after the hurricanes. • Drought and food insecurity: The 2014–2018 drought was highly linked to the positive phase of ENSO, for which there are international warnings and monitoring systems in place. However, Honduras at the time did not have a drought EWS. For food security, FEWS NET monitors the food insecurity conditions and FAO’s Agriculture Stress Index System . And there was not a system to monitor the impacts of the droughts, such as malnutrition rates, unemployment, migration etc. • Violence: There were no official early warning systems in place to monitor the various types of urban and rural violence in Honduras, although several NGOs and the Government do monitor acts of violence and report these statistics. In 2020, when inter-gang violence spiked during the lockdown, and extortion drastically increased after restrictions were lifted, several 80 agencies raised alarms in the media. Information sources for violence include the Observatorio Nacional de la Violencia de UNAH (UNAH - IUDPAS, 2019). Information on territorial control and rural trafficking routes is available, although this is highly dynamic information and will likely be more qualitative and highly sensitive information. • Migration: The International Organization for Migration (IOM) tracks and publishes information on international border crossings, people returned from Mexico, and to a lesser extent internal migration through its Displacement Tracking Matrix (DTM) platform. Especially internal migration remains a gap in relation to understanding dynamic population numbers in at-risk areas. • Following the early warning early action approach by the Honduras Red Cross, these are some vulnerability indicators related to root causes and dynamic pressures that could be used as part of an impact-based forecasting system: human development index, GINI coefficient, dependency rate by age, unemployment, population with disabilities, and a food insecurity index. • Furthermore, DRM capacity could be captured by assessing the presence of disaster risk management and emergency plans, municipal development index, access to electricity, percentage of households with internet access, number of mobile phone users, road density, access to improved sanitation services and literacy rate (HRC, 2021). • For exposure, OpenStreetMap could also play a key role for more granular exposure data. Annex 6: Disaster risk reduction recommendations • Structural reform of Government disaster management governance, including COPECO, to ensure national-, regional- and municipal-level coordination is in place for future crises. This includes more rigorous checks on public spending, improving communication, and staff training on disaster risk management procedures. • Prevention of future flood defence breaches and overtopping of barriers. Further improvement of flood barriers and drainage and ensuring continued maintenance of these structures would be critical to limit flooding impacts. Similar interventions after Hurricane Mitch contributed to a reduction in mortality from drowning and landslides during Eta and Iota. Such interventions should be managed on a basin scale. • To limit the sedimentation of rivers, land erosion, and excessive runoff in case of heavy rainfall events, reforestation in upstream areas and improved land management practices are critical. • Limiting settlement in low-lying areas close to rivers, on steep unstable slopes and in floodplains is crucial to reducing the number of people exposed to flooding and landslides. This requires the enforcement and implementation of existing laws (SINAGER), zoning plans and urban planning guidance, and addressing corruption and limited control of housing projects in at-risk areas. • To reduce the structural vulnerability of houses and other infrastructure, improvements in building quality through support for ongoing reforms (such as the Covivenda programme) and enforcement of quality standards, especially for low-income areas. • Streamlining national shelter regulations to ensure these adhere to international SPHERE standards by supporting municipalities in the implementation of their mandate for shelter. This includes regular updates of information on shelter location, ensuring there are enough shelters for the population at risk, that shelters are set up with adequate WASH facilities and that there are staff available to manage sites, among other things. This would contribute to a 81 reduction in the risks of the spread of disease, GBV and other adverse impacts arising from poorly managed shelter in future crises. • Increasing understanding of appropriate preparedness actions among disaster response agencies, emphasizing the importance of preparedness and early action. There were various examples shared where risks were underestimated by organizations, leading to loss of equipment, insufficient preparation or delays in humanitarian support to affected communities. • Co-production of DRR and climate change adaptation interventions. Especially for the predicted drying trend and intensification of storms. Event-based storylines for future planning that consider compound risk analysis like the one presented in this retrospective analysis can be instrumental to define actionable risk reduction measures, through “stress testing” projects, communities and strategies based on the experience before, during and after Eta and Iota. Annex 7: Data sources and additional outputs for mapping in this study 82 83 A6. Table 1. Data sources of INFORM Index Cod Name Description Dat File Sour Source Link e e For ce mat A1 honduras_pop_ Population Count & 202 tiff WO https://data.humdata.org/datase count, density Density 0 RLD t/worldpop-population-counts- POP for-honduras A2 water_bodies 201 shp Ries https://data.humdata.org/datase 8 gos t/honduras-water-bodies y Des arrol lo A3 livelihood_zone Most Predominant 201 shp ICA https://data.humdata.org/datase s Livelihood Zones 5 FEW t/wfp-geonode-ica-honduras- S most-predominant-livelihood- zones A4 roads 202 shp HOT https://data.humdata.org/datase 0 OS t/hotosm_hnd_roads M A5 waterways 202 shp HOT https://data.humdata.org/datase 0 OS t/hotosm_hnd_waterways M A6 airports 202 shp HOT https://data.humdata.org/datase 0 OS t/hotosm_hnd_airports M A7 health_facilities 202 shp HOT https://data.humdata.org/datase 2 OS t/honduras-healthsites M A8 sea_ports 202 shp HOT https://data.humdata.org/datase 0 OS t/hotosm_hnd_sea_ports M A9 Administrative 202 shp OCH https://data.humdata.org/datase boundaries 0 A t/cod-ab-hnd A10 wind_map wind density & speed 202 tiff DTU https://globalwindatlas.info/area 1 /Honduras A11 terrain tarrain map (no DEM) 202 tiff Wor https://worldclim.org/data/world 1 ldCli clim21.html# m/N ASA 84 A12 land cover ESRI land cover map 202 tiff ESRI https://www.arcgis.com/apps/ins 1 tant/media/index.html?appid=fc9 2d38533d440078f17678ebc20e8 e2 A13 climate global climate 201 tiff worl https://worldclim.org/data/world variables data 7 dcli clim21.html m A14 OSM_all 202 shp OS https://download.geofabrik.de/c 2 M entral-america/honduras.html H1 landdegradatio ICA landdegradation 201 shp WFP https://data.humdata.org/datase n_geonode_ma 7 t/wfp-geonode-ica-honduras- y2017 land-degradation H2 natural_shocks natural shock (floods, 202 shp UNE https://data.humdata.org/datase landslides and 0 P t/wfp-geonode-ica-honduras- droughts) risk - by natural-shock-risk second-level administrative area - H3 flood prone 201 shp UND https://data.humdata.org/datase areas 8 P t/honduras-other H4 Flood Extent Satellite detected 202 shp UNO https://data.humdata.org/datase Novembre 2020 water extents 0 SAT/ t/satellite-detected-water- between 13 & 17 NOA extents-between-13-17- November 2020 in A november-2020-in-honduras Honduras H5 landslide_risk 202 shp WFP https://data.humdata.org/datase 0 t/wfp-geonode-ica-honduras- landslide-risk H6 flood_risk 202 shp WFP https://data.humdata.org/datase 0 t/wfp-geonode-ica-honduras- flood-risk H7 Cyclone_tracks_ 202 shp EC- https://www.gdacs.org/resources ETA 0 JRC .aspx?eventid=1000738&episodei d=52&eventtype=TC&sourceid=N OAA H8 Cyclone_tracks_ 202 shp EC- https://www.gdacs.org/resources IOTA JRC .aspx?eventid=1000743&episodei d=21&eventtype=TC&sourceid=N OAA D1 drought_risk 202 shp WFP https://data.humdata.org/datase 0 t/wfp-geonode-ica-honduras- drought-risk 85 C1 ACLED 202 shp ACL https://acleddata.com/#/dashbo 1 ED ard V2 economic_insta information about 202 shp WFP https://data.humdata.org/datase bility the economic 0 t/wfp-geonode-ica-honduras- fragility levels - by economic-fragility second-level administrative area - estimated during the Integrated Context Analysis (ICA) run in Honduras in 2017 A6. Table 2 List of open-source geospatial data sets 86 A6. Table 3– Disaster timeline sources 87 A6. Figure 1. Rivers & waterbodies Comment map Rivers & waterbodies: This map shows major rivers, waterbodies and streams across Honduras. 88 A6. Figure 2. Terrain Model Honduras. Source: Map: Scholz (2022); Data: WorldClim (2017). Comment terrain map: The terrain map of Honduras shows that the country has very low elevation at, or close to, sea level in the south, north and especially north-east. Most area of the department Gracias a Dios lies at low elevation. The centre and west of Honduras is mountainous with elevations up to 2726 metres above sea level. A6. Figure 3. Population Density Honduras. Source: Map: Scholz (2022); Data: WORLD POP (2020). Comment map Population Density: The map of population density in Honduras (2020) shows that the country is most densely populated in the west and south of the country. The departments Cortés and Choluteca, and the municipality Districto Central have the highest population densities. Meanwhile, the east of the country is overall sparsely populated. The departments Gracias a Dios, Olancho and the eastern area of Colón have the lowest population densities. 89 A6. Figure 4. Administrative Boundaries of Honduras: 18 Departments (administrative level 1) and 298 Municipalities (administrative level 2). Source: Map: Scholz (2022); Data: OCHA (2020). A6. Table 4 contains the identified departments and municipalities facing compounding hazards as well as priority areas of geographic locations with high population density. Most impacted Most impacted Most people impacted within areas departments municipalities (disasters & conflict weighted with Population Density) (admin 1) by (admin 2) 2018– disasters 2018– 2020 (disasters & 2020 (>=5 conflict incidents) disasters) Santa Bárbara Santa Bárbara City of Santa Bárbara & surroundings Cortés Puerto Cortés District of San Manuel Omoa City of San Pedro Sula & surroundings Choloma Eastern San Pedro Sula district San Pedro Sula Central Choloma district La Lima Omoa city & coastal area Santa Cruz de Yojoa Puerto Cortés city & coastal area Villanueva El Progreso city & surroundings Potrerillos Agua Blanca Sur city & surroundings 90 Santa Cruz de Yojoa Santa Cruz de Yojpa & surroundings San Manuel Pena Blañca city & surroundings Villanueva city & surroundings El Milagro city & surroundings Posterillos city & surroundings Olancho Catacamas Catacama city & surroundings Juticalpa Juticalpa city & surroundings San Francisca de Becerra city & surroundings Atlántida Tela Coastal area of Tela Jutiapa Coastal area of La Ceiba La Ceiba Atlántida city Yoro Olanchito Coyoles city Yoro Olanchito city Victoria El Progreso Comayagua Comayagua Siguatepeque city & surroundings Siguatepeque Comayagua city & surroundings Choluteca Choluteca Choluteca city & surroundings Agua Caliente city & surroundings Francisco Distrito Central Tegucigalpa city & surroundings Morazán Western, southern and eastern area of Distrito Central A6. Table 4. Summary of identified departments and municipalities facing compounding hazards as well as priority areas of geographic locations with high population density. 91 A6. Figure 5. Municipalities impacted by compound hazards in northwest Honduras. 92