Multi‑hazard early warning systems (MHEWS) dominate the narrative-from global frameworks to national plans. Yet what “multi‑hazard” means when applied practically to an EWS remains fuzzy, and realities on the ground lag behind: most systems still do not account for how hazards and risks interact. This blog post distils insights from our recent perspective to move from rhetoric to practice.
The problem: hazards and risks don’t queue up
Communities at risk rarely face hazards one by one. Wildfires can follow droughts, landslides follow earthquakes, and a cold wave can hit during a flood. Multi‑hazards take several forms: triggering (one hazard sets off another), amplification (one hazard changes the likelihood or severity of another), compound (impacts coincide in time and space), and consecutive (back‑to‑back events before recovery is complete).
Yet most early warning systems (EWS) were built for single hazards and rarely account for these interrelationships. The result can be confusing messages—or worse, contradictory advice—precisely when people need a clear course of action.
What we need: people‑centred, genuinely multi‑hazard EWS
A MHEWS recognises that because hazards may occur independently, simultaneously, in sequence, or cumulatively over time, a multi-hazard EWS requires consideration and addressing those interactions within each EWS component and as a holistic approach across the system.
In addition, for it to be people‑centred, a MHEWS must also explicitly account for differentiated vulnerabilities and capacities and how age, gender, disability, livelihood, language, mobility, social status, and location shape whether people can receive, trust, and act on a warning. It should also reflect how vulnerability and exposure change across an unfolding event and different hazards, and tailor lead times, dissemination channels, and early actions to what different groups need and can actually do.
Four insights from multi-hazards to reframe how we warn
Hazards interact – warnings must too
Triggering, compounding, amplifying, and consecutive events are normal, not exceptions. If agencies issue separate alerts without coordination, people can receive conflicting guidance.
Vulnerability is dynamic
A first shock (e.g., flood) can reduce a family’s ability to act on a second (e.g., cold wave). Effective warnings account for how vulnerability, exposure, and options evolve during multi‑hazard events.
Multi‑hazard impacts are bigger
Interacting hazards often drive higher losses than isolated ones. Treating them separately underestimates risk and delays the right actions.
Multi‑risk is the real world
People navigate overlapping challenges. Whilst challenging and complex, MHEWS must link risk knowledge, monitoring, communication, and preparedness across sectors so that people receive accurate and actionable information that reflects the reality they face.
| Interrelationship type | Description | Example implication for MHEWS |
| Triggering | One hazard triggers one or more other hazards. For instance, rainfall-triggered landslides in Nepal. | Scenario of an earthquake-triggered tsunami: people advised to leave their houses due to an earthquake, then at a higher risk of a tsunami because of being outside. |
| Amplification | One hazard changes environmental parameters, resulting in an increased probability of another hazard occurring. For instance, heatwaves increase the probability of wildfires in East Africa. | Warning thresholds need to be updated regularly, to account for changes in hazard probabilities. For instance, after a wildfire, there is an increased risk of flooding as surface runoff increases – flood warning thresholds need to be updated with the information from other hazard types. |
| Compound | Hazards and their impacts coincide in time and space. They can be resulting from the same primary event or driver or have no underlying interrelationship. For instance, compound flooding in India caused by heavy rainfall and water level rise in Southern Kerala in India, or COVID-19, Cyclone Amphan, and monsoon floods in Bangladesh in 2020. | The impact of the hazard is likely to be affected by two hazards occurring simultaneously. This has a direct impact on early action and communication of risks – e.g., whether to shelter in place (to avoid exposure to COVID-19) or to evacuate (to escape flood waters), whether to shelter in the basement (to avoid wind damage from storms or tornadoes) or on the top floor (to avoid flooding). If there are dual warning systems for different hazard types, people may be receiving conflicting alerts from different sources advising to do different early actions. |
| Consecutive | One or more hazards occur in succession, with their direct impacts overlapping in the same area, while recovery from the initial event is still underway. For instance, consecutive Cyclones Idai and Kenneth in Malawi, Mozambique, and Zimbabwe in 2019. | Due to losses incurred by an initial event, people’s capacities to access resources and prepare for a consecutive event are diminished. Disaster-related migration and/or displacement from the primary hazard can also place people at greater risks to subsequent hazards. |
Five practical entry points
Assess honestly
Critically assess where the EWS is on the MHEWS spectrum – in other words, does it cover a single‑hazard, multiple single hazards, connected single hazards, or full multi‑hazards (accounting for interactions and changing vulnerabilities). Take action to make progress forward on the spectrum; for example, move on from a single-hazard system by picking one or two high‑value interactions (e.g., wildfire–flood, heatwave–drought, rainfall–landslides) and design joint triggers and early actions.
Start with what you have – connect it
If you operate good single‑hazard systems (e.g., cyclones, floods, landslides), link their data, thresholds, and messaging. Where feasible, move toward impact‑based forecasting that translates “what the weather will be” into “what it will do” for people. In Bangladesh, cyclone alerts are paired with tidal‑surge information and rainfall alerts warn of possible rainfall‑triggered landslides; and in Nepal, impact‑based forecasts are being developed for rainfall‑triggered landslides.
Coordinate across sectors and agencies
Agree who leads what, share data by default, and pre‑clear joint messages. In many countries, different agencies are responsible for different hazards (e.g., hydrometeorology vs. geology), so collaboration and clear procedures are essential. For example, in Nepal, a mandate for EWS is assigned to the Department of Hydrology and Meteorology, who have expertise on floods and weather-related hazards, whilst landslide expertise, essential for landslide EWS resides in the Department of Mines and Geology.
Make equity and inclusion central
Co‑design messages and early actions with those most at risk, including women, older people, youth, people with disabilities, migrants, seasonal workers, and indigenous communities. Ask what is actionable for them at different lead times, then tailor channels (e.g., voice, radio, SMS, in‑person) and content accordingly.
Engage communities as co‑producers, not just recipients
Community brigades, citizen science, and local resilience agents can feed real‑time observations, validate thresholds, and deliver last‑mile alerts when infrastructure fails. In Peru’s Rímac basin, communities are engaged in localised monitoring networks and community civil defence brigades support orderly evacuations.
What progress looks like
Building MHEWS will look different in each place and should reflect local needs and risks. Our paper suggests a few starting points: practical examples to guide early steps toward MHEWS:
- Risk knowledge that reflects people’s priorities. Combine local knowledge with data to identify relevant hazards and interactions, and who is most at risk of what.
- Monitoring that serves decisions. Reuse stations and channels (e.g., flood gauges for heat/cold monitoring). Fill gaps with community observations.
- Communication that’s coherent and clear. One authoritative, pre‑agreed message during compound events; plain language about impacts and uncertainty.
- Preparedness that avoids “asynergies”. Plan for actions that help across hazards (e.g., shelters doubling as cooling centres) and spot trade‑offs early (e.g., water allocation in droughts vs. firefighting capacity).
A call to act differently
“Multi‑hazard” does not require perfect science before we move. It asks us to connect what already exists, centre those most at risk, and design alerts and actions for the way risk actually unfolds. If we do that, link by link, protocol by protocol, we will turn today’s ambition into tomorrow’s saved lives and protected livelihoods.
This blog post, written by Robert Šakić Trogrlić and Mirianna Budimir, shares key insights from a wider team of Practical Action staff across Latin America, Africa, and Asia, based on the paper Opportunities and Challenges for People-Centered Multi-Hazard Early Warning Systems: Perspectives from the Global South.
This blog post was published by PreventionWeb on 8th September 2025. You can view the original here.
Comments