06. Determine activation trigger

Summary

Forecasts and triggers are among key components of anticipatory action that will guide the National Society, the community or other practitioners to when and where early actions need to be implemented. For s/EAPs, when a pre-agreed threshold of a forecast is reached, the National Society activates the s/EAP and funds will be automatically dispersed.

Anticipatory action is designed for those events that are predicted to have a severe impact exceeding the coping capacities of the people. Therefore, it is not enough to know what the forecasted windspeed will be. Instead, you must know whether the storm will cause impact. For this reason, as impact-based forecasts are not usually available the development of the trigger requires an understanding of past impacts that can be related to a certain hazard threshold, e.g. windspeed.

The methodology elaborated below can be applied and adapted for a variety of hazards. It is derived from work on weather-related hazards, such as floods, storms and droughts, but it can also be adapted to non-weather hazards, such as epidemics or locusts, if appropriate.

This chapter will guide you through the trigger development process from a technical perspective with a focus on trigger development for EAPs. It will also advise you on technical cooperation and share examples and good practices. However, many steps can be taken for the other approaches in a lighter touch depending on your available time and resources. You find guidance for this throughout the text and the specific guidance at the end of the Manual.

Please note that the development of the trigger often runs in parallel with the selection of the early actions, as the development of the s/EAP is an iterative process. For example, if the action is “evacuation”, you would select a different triggering forecast than if the action is “spread awareness about cholera risk”. Thus, some steps from chapter 7: Select your early actions are needed to advance with the trigger development.

In the video below, Catalina Jaime, expert on anticipatory action, climate and conflict, explains how to build upon a feasibility study, and complete a risk analysis.

Step 1: Select hazard

If not done during the scoping study, you need to decide which hazard you will address with your anticipatory action system. In the s/EAP you will be asked to provide a brief rationale about the selection of the hazard and when it has caused various humanitarian impacts in the past (Section 3.1 in full EAP). This should include:

  • A table that outlines the most severe events of this hazard in the past (see a simple template in the toolbox).
  • A written description of what the impacts of this hazard were, who was affected and why, how many people were affected, and what were the main sectors affected

Please refer to chapter 5: Collection risk, early action and impact information to learn more about the data collection of these information.

Step 2: Understand what needs to go in a trigger

The trigger will allow the National Society to know when and where to act, regardless of whether the hazard is meteorological, epidemiological or socio-political in nature. For the s/EAPs, please verify the quality criteria for the trigger and the s/EAP activation processes that are linked in the toolbox below. For most anticipatory action approaches, you need to include three components of an operational trigger system:

Intervention map: When do you need it?

Depending on the anticipatory action approach, you may already know where to act once the trigger is reached, but in other cases, this is not yet clear. For example, when working with a specific community, you can develop a trigger and an anticipatory action plan tailored to that community. In such cases, an intervention map is not necessary.

For s/EAPs, however, the situation is different. They usually cover a larger geographic area, and the intervention area is not defined in advance. Only once the forecast is available does it become clear which communities are likely to be affected. This information is then often combined with vulnerability data to prioritise the communities expected to face the greatest impact.

Step 3: Establish coordination mechanism

In order to ensure sustainability, ownership and alignment of the trigger, it should ideally be embedded in the national early warning and alert system. At the very least, key actors should be involved in or endorse the trigger development process and its outcomes. Ideally, as trigger development can be a very technical process, you cooperate with meteorological services, disaster management authorities, and other experts and institutions that work in the field of interest. This also enables other government agencies to participate in anticipatory action. To begin with, you can think about the following aspects that might also shape your advocacy strategy (please refer to chapter 2: Engage your stakeholders).

Mandates
Who has the mandate to issue forecasts and early warning messages? Ensure that you are in line with the official regulations and procedures. For extreme weather events, for example, the meteorological service usually has the mandate to issue forecasts and early warning messages together with disaster management authorities. For epidemic outbreaks, however, the ministry of health might be the relevant authority for surveillance and / or early warning. Regardless of the hazard in question, integrating or the very least aligning the trigger with relevant mandates is important to avoid confusion, which could happen if there were many actors issuing their own alerts.

Existing initiatives

In many cases, the national meteorological agency is already involved in initiatives around impact-based forecasting or anticipatory action. For example, in Mozambique the government partnered with the World Food Programme (WFP) to develop harmonized drought triggers that are owned by the government. If such early warning mechanisms already exist, you should tap into those to avoid a duplication of work. Perhaps there are existing agreements, MoUs, or trigger activation systems already established.

On the regional level, technical working groups also shape the anticipatory action landscape and thus trigger development. There are several examples in Africa, but also in Latin America and Asia Pacific. Please find an overview of the regional technical working groups here. These technical working groups can often provide overviews of existing initiatives that are relevant to the trigger development in the context of interest.

On local level there might already be a community early warning system in place that can be built upon and adapted for anticipatory action or municipal authorities might have plans in place that include elements of anticipatory action.

Technical standards in the Asia-Pacific region

As anticipatory action is gaining momentum in Asia-Pacific, there is a growing demand for coherent terminology and approaches to building an anticipatory action system. Recognizing the need, the Asia-Pacific Technical Working Group on Anticipatory Action (TWG AA) has collectively developed Technical Standards on Anticipatory Action in Asia and the Pacific through a consultative process with its members. The document draws upon the definition and three key building blocks set out in the ASEAN Framework on Anticipatory Action in Disaster Management. You can access the document via the toolbox.

Guiding questions to establish a coordination mechanism:

  • Who are the key actors issuing forecasts and early warning messages?
  • Are there existing initiatives on anticipatory action, such as technical working groups?
  • Are there existing agreements or risk information systems that could be used for any of the three parts of the trigger (trigger statement, intervention selection, and trigger monitoring and activation processes)?
  • Is there political will within the government to own and maintain the trigger monitoring and activation processes?
Step 4: Define hazard variables to monitor based on impacts and early actions

In order to understand which forecast(s) would be most suitable for your trigger, you need to understand the impact it should forecast. For example, when talking about impacts of cyclones you need to understand what has caused the impact: the winds, the rains or the storm surge? Hence, here you need to review the work that you have done on impact analysis and potential early actions. Please refer to the guidance in chapter 5: Collect risk, early action and impact data and chapter 7: Select early actions regarding the prioritization of impacts and early actions.

Hazard Cascading hazard Potential primary and secondary impacts Variable and drivers
Heavy rains Riverine Flood - Landslides Damaged houses - Damaged crops - Livestock dead - Cholera outbreak Water level - Flood extent
Cyclone Damaged houses - Damaged crops Wind speed - Strong rains - Cyclone-induced floods - Storm surge
Drought Agricultural droughts - epidemics Death of livestock due to insufficient pasture - Crop failure Vegetation condition - Rainfall - ENSO

Guiding questions to define hazard variable

  • Which impacts were prioritised?
  • What has caused the impact?
  • What variable can be used to best forecast the prioritised impact?
  • How much lead time is needed to implement the early actions? What early actions would be feasible in a given lead time?
  • Can this variable be forecasted with the needed lead time?

Note that a trigger system does not necessarily need to rest on a single variable as described more extensively under step 10 in this chapter. The trigger for the cholera EAP in Cameroon, for example, rests on a combination of climatic indicators that are known to impact WASH infrastructure and thus make transmission more likely on the one hand, and an indicator for suspected or confirmed cholera cases in relevant districts on the other that is based on surveillance data. Keep this in mind when selecting relevant variable(s), especially for non-weather-related hazards.

Step 5: Compile a menu of forecasts

In order for anticipatory action to allocate funding when a trigger is met, there must be a certain probability of the event. To safeguard this, it is crucial to select the best-suited forecasts. For weather-related hazards, this involves meteorological forecasts. For non-weather hazards, this could involve disease surveillance systems, early outbreak detection models, displacement monitoring tools, or qualitative alerts. Regardless of the hazard in question, for s/EAPs, but particularly full EAPs, an analysis of forecast and / or data verification, type, reliability, lead times, and sources of data for forecasts should be presented as an inventory, to allow an informed decision about which one to use. If the disaster is already imminent and you want to use an imminent DREF, you would most likely rely on a national early warning or similar information which has been already issued.

Note that forecasts do not need to be analysed or calculated by the National Society but can be jointly worked on or led by  hydro-meteorological services, research institutions, or other experts.

Guiding questions for the forecast selection:

  • Which agency produces relevant forecasts on local, national, regional or global level? Are there any community early warning systems that could be used or established?
  • What kind(s) of forecast(s) are produced or how is the forecast produced? Tip: Choices include observed data (e.g. gauged precipitation), statistical forecasts (e.g. extrapolation of an upstream river flow to a downstream location, or an index based on El Niño sea-surface temperatures), and dynamic models (e.g. numerical weather-prediction systems and large-scale hydrological forecasting models). Check the toolbox for more information.
  • What outcome is the forecast forecasting? E.g. seasonal total rainfall, flood levels, windspeeds.
  • Is the forecast deterministic (showing a single outcome without conveying potential error and uncertainty) or probabilistic (showing the probabilities of one or more discrete outcomes or categories) or issued as an interval (showing an explicit upper and lower limit between which a value is expected to occur)?
  • What is the spatial (e.g. what is the level of detail?) and temporal (e.g. how often is the forecast produced in a month?) resolution of the forecast?
  • What is the lead time for each forecast, e.g. what is the time between forecast issuance and the shock?
  • What regions are covered by the forecasts?
  • What is the skill (the quality) of the forecast? How has the skill been assessed (e.g. skill at a specific location, skill at predicting extreme events)?
  • Is the forecast available throughout the year? Does the accuracy / skill of the forecast change over time?

The menu of forecast required in the full EAP (section 4.2) requests to provide the following information about the available forecasts and requests to include an explanation about which forecast will be used and why. Especially for non-weather-related hazards, this menu might include relevant non-meteorological monitoring or surveillance systems (e.g. health or protection sector dashboards, and Internally Displaced People (IDP) tracking systems).

  • Name of forecast
  • Source
  • Lead time
  • False alarm ratio and missed events
  • Number of times the forecast has been issued for this hazard in the last 10 years

You can download a template for the menu of forecast table and find a more comprehensive guide on forecast verification or skill analysis in the toolbox below.

Menu of forecast and skill analysis for the simplified EAP

For the sEAP a less thorough assessment can be done. The rationale is that perhaps there is less data available to assess the forecast skill which is particularly true for non-weather-related hazards. However, if data is available, information should be included as you want to avoid false alarms or missed events that will jeopardise the reputation of the National Society but also the mechanisms of anticipatory action.

Menu of forecast and skill analysis for community-based approaches

For community-based or very localized approaches, there may not be the capacity to carry out an in-depth forecast skill assessment. However, you can still compile an overview of the forecasts and warning information available, such as national forecasts, community-based early warning systems, and traditional or Indigenous forecasting methods.

Even at community level, you will need to select a forecast or signal to prompt early action. This selection process may not be as scientifically rigorous as in other contexts, but what matters is that the community trusts the information enough to act on it. For that reason, validating the chosen forecast or signal with the community is essential.

If there is significant uncertainty about forecast reliability, consider focusing on no-regret actions.

Step 6: Set impact threshold

In whatever anticipatory action approach you selected, you want to address hazards of a specific magnitude. For example, the purpose of the DREF is to fund humanitarian operations by National Societies. Hence, the anticipatory pillar of the DREF provides funding for anticipatory action for events of a magnitude that can be linked to significant humanitarian impact in the past. For your community-based approach this threshold might be lower or the government might also set different thresholds. Still, it is helpful to collect information on the previous impacts of the selected hazard in the country or context to determine and justify the selected impact level. Please revisit chapter 5: Collect risk, early action and impact information, if you still need to collect impact data.

Guiding questions

  • From your impact assessment, which years show humanitarian impact, e.g. because a significant number of people was affected, a significant number of houses destroyed?
  • In which years did the National Society respond to a hazard, e.g. with a DREF allocation or with other humanitarian funding?
  • Which years or events were perceived as disastrous by communities, branches, experts or other stakeholders?
  • When was a disaster declared for the selected hazard?

The discussion should result in a decision at what impact level you would want to activate your plan. For example, you could use reference events stating that you are targeting events like a cyclone with windspeeds greater than a certain amount, which happened in the past in your country.

Note that sEAPs target the same type of events: those that have caused humanitarian impact in the past. However, the magnitude of the event will differ between an EAP and a sEAP. Exceptions are cases where only very little data is available.

The full EAP (section 4.3) asks you to do the following:

  • explain why you selected a certain magnitude of the event,
  • how your impact level was defined,
  • how much impact can be expected based on the strength of the event,
  • indicate to which return period the selected impact level corresponds

Why do EAPs target only events that have caused humanitarian impacts in the past?

Anticipatory action is not the silver bullet to solve all challenges in disaster risk management or address skyrocketing humanitarian needs across the globe. Anticipatory action aims at managing residual risk that cannot be addressed by long term investments in disaster risk reduction. For example, in a flood prone area you would try to build a dyke or put houses on poles if the houses are flooded on an annual or bi-annual basis rather than setting up an anticipatory action system for those recurrent events. On the other hand, you are unlikely to have enough funding to build a dyke so high that it will also reduce the risk of a major flood. Anticipatory action can be critical to evacuate people and their livelihoods before that kind of major flood.

Setting impact magnitudes

The image can help in explaining which events we target with anticipatory action. You can see that people can cope with recurrent floods (green line) by building their houses on poles. Stronger events (orange and red lines) that occur less often will then also have impact and require action, e.g. evacuation of people and their assets.

Step 7: Link impact and hazard magnitude

After having defined what magnitude of impact you want to target, you need to link the impact magnitude with the hazard magnitude, e.g. estimating what windspeed would cause houses to be destroyed. Depending on your availability of data (both impact and hazard), you can plot historical events on an x-y axis, with the hazard magnitude (e.g. windspeeds) on the x-axis, the impacts (e.g. houses destroyed) on the y-axis. This is called a “hazard-impact curve” and can help you decide at what level of hazard would be important to use as your trigger, because you know it causes substantial impact.

Approaches to hazard-impact curves

Example 1: Philippines Floods – Simple - Impact data
For floods in Philippines, the analysis of a lot of impact data is taken into account depending on the flood level, such as the price drop of rice crops depending on their reproductive stage and the percentage of crop loss on municipal level that is considered critical in a five year return period event. The Philippines are employing a very impact and data-driven approach.
Example 2: Ecuador Volcanic Ash - Expert estimation
In Ecuador, the impact of volcanic ash is categorized into three main areas: crop loss, damage to livestock, and health effects on people. Estimated impacts for each category are determined by the quantity of volcanic ash present in the region. As a result, three distinct and practical levels of impact have been clearly delineated, providing a structured framework of thresholds. It is important to note that this model relies heavily on expert opinions, with less reliance on empirical and experimental data.
Example 3: Bangladesh Cyclone – Vulnerability data crossed with historical data
In Bangladesh the impact on houses is prioritized, therefore several factors are combined, such as the distance of houses from the coast, the proportion of kutcha and semi-pucca houses and the wind reduction factor to estimate the damage of a certain wind-speed to these houses. This information is triangulated with historical data about house destruction. When a storm approaches, the forecasted wind speed is transformed into an exposure map. This map, combined with impact curves, estimates the percentage of houses likely to be affected by the forecasted storm. The result is a map highlighting the percentage of houses at risk of destruction within each union in the exposed district. Priority for intervention is assigned to unions where more than 25% of rural houses are potentially damaged.
Example 4: Mongolia Dzuds – Prior Events
In Mongolia not a lot of data is available on both the forecast and impact aspects, making it difficult to draw definite conclusions about the level of impact. The action threshold defined in the trigger section is derived from analyzing the limited data available for the past few years. It represents the best estimate to identify years when the herder population and provincial government may find it challenging to safeguard herder livelihoods from substantial losses.
Step 8: Define and justify trigger threshold and forecast variable

After having analysed the link between impact and hazard magnitude you should be able to determine and justify the level of impact and the corresponding hazard magnitude at which you want to trigger your plan or s/EAP. Hence, the output of this step should be a threshold of a given forecast variable and its justification. For the full EAP two more aspects are specifically mentioned in the criteria which might also be useful to think about in the other approaches:

  • Frequency of the threshold being reached: Please show in which years (if the data allows) you would have reached the trigger and what is the frequency of the trigger being reached
  • Humanitarian impacts: Rather than looking at specific return periods, you will need to demonstrate that your threshold is linked to events, which have caused significant humanitarian impacts in the past (see below).

Dynamic or flexible triggers

Triggers are very useful when pre-agreed. Moreover, most of the triggers in s/EAPs are static which means the threshold is set on one specific value, e.g. windspeed of 120km/h. However, what if the reality doesn’t follow your plan, e.g. the very same region is hit by a cyclone again, we have consecutive drought years or communities are more vulnerable because of conflict outbreaks and even a small flood will have huge impacts? This could be addressed by more flexible or dynamic triggers that take into account changing vulnerabilities and contexts.

Impact-based forecasting

The idea of impact-based forecasting (IBF) is to predict what the weather will do and not what the weather will be like. However, IBF requires substantially more data than only weather forecasts, e.g. meteorological data in combination with risk information.

Please find a guide to impact-based forecasting in the toolbox.

Step 9: Define stop mechanism (if applicable) 

For anticipatory action plans with a lead time longer than three days, a stop mechanism may be useful to be considered and is required for s/EAPs. A stop mechanism stops the activation of early action activities in case there is a significant change of the forecast, and the threshold is no longer reached, or it is still met but in a different location. Hence, the stop mechanism should include the description of what the National Society will do if, according to the forecast, the hazard is projected to change in intensity or location within the last days before the event. That means that also actions need to be planned accordingly, so that they can be stopped at the point of the stop mechanism.

Bangladesh Flood EAP

Bangladesh Flood EAP

In the trigger statement of the Bangladesh flood EAP, you can see that they introduced a 10-day pre-activation trigger in which they start to pre-register families. However, only when the 5-day activation trigger is met, the early action is implemented. If the 5-day trigger is not reached or confirmed, the activation and thus the early actions will be stopped.

Step 10: Develop a trigger statement

The trigger statement is a key component of your pre agreed anticipatory action plan (section 4.1 in the full EAP) and it should exactly state when your plan will be triggered. The trigger statement should be clearly understood by all stakeholders involved in order to avoid confusion about when a trigger is reached.

A trigger statement typically includes the following ingredients:

  • Forecast used and forecast source: The forecast used and forecast source should clearly state which forecast, e.g. a certain rainfall forecast is being used and from what source, e.g. national meteorological service
  • Variable: the variable should clearly state which variable of the forecast is being used. Hence, this should be an indicator such as Standardised Precipitation Index (SPI3) or heat index, or water levels at a specific station.
  • Threshold of that parameter: Here you put the threshold that you have identified in the previous step. Make sure that your forecast provider is calculating this threshold, e.g. if the forecast is issued in mm of rainfall, but your trigger is based on the standardized precipitation index (SPI), who will calculate the SPI?
  • Lead time: The lead time defines when the conditions must be met for the anticipatory action to be effective. This helps ensure that interventions are timely and proactive.
  • Probability: If available, you should add the probability of the event to happen. For example, the trigger could be a forecasting showing 80% chance or greater of 30mm of rainfall or more.
  • Stop mechanism: If your plan has a stop mechanism, it should be mentioned in the trigger statement.
  • Monitoring system: You need to decide how your forecasts are monitored and who is responsible for this. We will elaborate this further in chapter 8.

Examples of trigger statements

  • Philippines

  • Mozambique

  • Niger

A complete trigger statement can look like this:

When the [Source of information] forecasts [Threshold of climate variable OR Threshold of impact] at [Lead time] and with a [measure of probability], actions will be taken. If [stop mechanism] occurs, activities will be paused. This system will be monitored closely by [Monitor] in the following way [System].

Staggered or phased triggers

In order to win valuable time, some National Societies have introduced phased or staggered triggers. That means that they have a pre-activation trigger and an activation trigger. The Bangladesh example above introduced a phased system to their flood EAP. They use a pre-activation trigger at day -10 to pre-register households and get some logistics ready, to then be ready to implement their actions, e.g. cash distributions at day -5. This gives them valuable time as without the readiness trigger, the actions could not be implemented in the short lead time.

These triggers also constitute important approaches to anticipatory action for non-weather hazards, since events like epidemic outbreaks or displacement crises are difficult to anticipate relying on a single (meteorological) variable. Instead, anticipatory action for epidemic outbreaks might rely on a combination of climatic indicators (e.g. rainfall levels for cholera) that are known to favour transmission as pre-activation triggers and surveillance indicators (e.g. suspected or confirmed cases in neighbouring districts) for the actual activation of the plan. Likewise, triggers for displacement might rest on border-crossing reports and a subsequent assessment of coping capacities in estimated host communities.

Note that funds are only allocated from the IFRC-DREF for an s/EAP once the activation trigger is met. If funds are needed during the phase between readiness or pre-activation trigger and activation trigger, the National Society has to find them or need to account for them in the annual readiness costs of the s/EAP.

Combination of forecasts for drought

In some cases, a combination of different forecasts and variables might be needed. In a Validation Committee meeting in 2023, the members suggested to combine different forecasts in drought triggers. The rationale is that rainfall forecasts alone will not suffice as they do not give any information on the impact (due to the time lag and also the timing of rainfall and its impacts on vegetation). Hence, rainfall forecasts should be complemented with measures of vegetation condition, food security or any other measure that reflects the prioritised impacts.

Combining information on different aspects and / or from different sources is also important for non-weather hazards. As outlined above, triggers for epidemic outbreaks or displacement might rest on a combination of indicators that are associated with their occurrence, such as monitoring and surveillance data, occupancy levels of reception or healthcare facilities, or political and / or weather-related drivers.

Step 11: Develop a mechanism that defines the intervention area

Once you receive information about the trigger being reached you should also know in which geographic regions the trigger is reached. If you work in a preselected community or area designated by the government, you activate your anticipatory action plan. For anticipatory action plans that cover larger areas like s/EAPs, you need to know how to prioritize the municipalities and where to intervene. Note that you and the National Society will probably not have the capacity to act in all municipalities and support everyone, hence you will need to prioritise in a transparent manner. The full EAP criteria state:

There is a map or a clear methodology that will tell the National Society where action should be taken based on a combination of vulnerability, exposure, and the forecast, when the EAP is activated based on the trigger model. (Section 4.4 in full EAP template)

Currently there are two main ways to define the intervention area:

Impact-based forecasting

In case you have an impact-based forecasting system, it should tell you which locations might be most affected. For example, if your National Society has the capacity to act in three municipalities you can select those three municipalities that are forecasted to be most affected.

Indicator-based risk assessment

If no impact-based forecast is available, you can use a regular weather forecast and overlay this with exposure and vulnerability information. This can serve as a proxy to predict where impacts will be highest. The simplest way to do this is to use an indicator-based risk assessment by using the vulnerability and exposure indicators that you have identified in your risk assessment (see chapter 5: collect risk, early action and impact information). This method is used in most EAPs so far.

There is a variety of methodologies to do indicator-based risk assessments. However, the most prominent one is probably the INFORM Risk Index methodology.

Automatization in QGIS

When an EAP activation is near, speed matters. An intervention map is essential for planning and communicating your response, but creating such maps usually takes hours and advanced GIS skills. With QGIS automation, the process can be reduced to minutes. QGIS is a free and open-source GIS used worldwide, supported by a strong community. Its automation features allow you to run predefined processes with a single click – no coding required. As long as your team has basic QGIS skills, you can create, adapt, and troubleshoot automated workflows. Unlike many tools tied to administrative boundaries, QGIS enables precise geospatial analysis, helping you target the areas most at risk.

Specialized QGIS Training for Automated Trigger Models

The IFRC Network QGIS Training Platform is the main resource for building QGIS skills in a humanitarian context. It includes a dedicated exercise track on developing automated trigger workflows, based on the Madagascar Cyclone EAP model. Documentation of the operational model is also available on the platform. 

Through this exercise track, practitioners learn how to: 

  • Build a model from scratch. 
  • Adapt existing models. 
  • Tailor outputs (e.g. generate spreadsheets). 
  • Apply predefined styles and map templates for rapid map production. 

The track is embedded within the wider training platform, giving easy access to video tutorials, additional exercises, and in-depth guidance on GIS analysis. You can find a link to the training platform below in the toolbox. 

There is no defined methodology to derive the intervention map. Hence, if you have another established methodology, please make sure that you explain it well in the EAP document. You can find examples from Madagascar and Somalia below.

Example 1: Cyclone trigger development in Madagascar

For the AA project in Madagascar, the activation trigger is set as follows: if the Météo Madagascar (DGM) forecast indicates landfall of a tropical cyclone with wind speeds above 118 km/h within the next 48–72 hours. 

Once this condition is met, an automated QGIS model defines the intervention area and estimates expected impacts. The model buffers the forecast cyclone track by 200 km and intersects it with multiple datasets: the exact locations of health and education facilities (OpenStreetMap), and a 1 × 1 km raster grid containing population (WorldPop), number of buildings (Google Buildings), and agricultural area (Copernicus Land Cover). 

The model outputs both geospatial layers and spreadsheets. Predefined style files and map templates allow users to quickly generate: 

  • An overview map highlighting the affected districts. 
  • Thematic maps for each sector, showing the exposed population, buildings, agricultural land, and health and education facilities. 

The model relies on one dynamic input – the forecast cyclone track – while all other datasets remain static. These static datasets are not always the most up-to-date or perfectly accurate, but they have a crucial advantage: they are freely available, cover the entire country, and even exist globally. This makes the workflow simple, fast, and reliable. 

Ground-truth data, such as detailed government surveys or field assessments, can provide higher accuracy and local relevance. However, these datasets are often costly, slow to collect, and only available for limited areas. Global datasets, by contrast, may be less precise at the local scale, but their consistency and coverage allow organizations to build models that are immediately operational and can be easily updated when new versions become available. 

Other trigger models can achieve higher precision by relying on advanced technologies or highly detailed datasets. Yet these approaches often come with trade-offs: higher costs, greater technical complexity, and reduced local ownership, as many organisations lack the resources or expertise to maintain and adapt such systems. The Madagascar cyclone model shows that a straightforward QGIS-based approach can remain both effective and sustainable within existing capacities. 

The QGIS trigger workflow for Madagascar can be found in the toolbox.

Example 2: Drought trigger development in Somalia

For the anticipatory action project in Somalia, a whole workflow was developed in the open source geoinformation software QGIS for the trigger which is owned and run by the Somali Red Crescent Society.

For the development of the Somaliland-Somalia Drought Trigger mechanism various data sources were thoroughly analysed. Finally, the main parameters chosen for the trigger, based on the historical impact assessment, are the twelve-month Standard Precipitation Index (SPI12) and the IPC acute food insecurity classification. The exact data used are the documented and forecasted SPI12 (source: ICPAC) and the forecasted IPC classification (8-month forecast, source: FEWSNET), that is used to calculate a population weighted index of food insecurity. The trigger thresholds for both components were optimized towards the most favourable proportion of hit rate and false alarm rate. The emerging thresholds were <-1 for the SPI12 and >=0,7 for the IPC-based index. The triggering is done on district level and per district just one trigger activation per year is possible.

This information of the threshold is then overlaid with an indicator-based risk assessment which identifies the most at risk districts. The QGIS trigger workflow for Somalia can be found in the toolbox.

Toolbox

Tools and templates

Impact-based forecasting

Forecast skill assessment

Drought related guidance

Non-weather hazards

Other key reads, documents and websites

Quiz

Chapter 6

1 / 5

What is the main function of a trigger system in the Forecast-based Financing (FbF) approach?

2 / 5

Why is it important to connect a forecasted hazard with its past impacts?

3 / 5

Which elements should be included in an operational trigger system? (Select all that apply)

4 / 5

What data should be analyzed when defining hazard variables for a trigger system? (Select all that apply)

5 / 5

Which of the following statements are true?

Your score is

The average score is 53%

0%