Previous research has found that users fail to comprehend flood hazard warning messages as well as the possible impacts of the forecast events. A proposed way to improve understanding and uptake is the implementation of impactbased warning services. However, even though extensive qualitative research has been done on the provision of these impact-based warnings (IBWs) for different types of hazards and extreme events, there is still little scientific evidence that the additional information on impacts in flood warnings positively affects decisions made by the general public. The research reported here explores whether including messages of potential impacts increased the general public's likelihood of making a protective decision when provided with a flood warning. Nine hypothetical flood scenarios were presented to participants who were then asked to score their likelihood of making a specific protective decision. Participants were either presented IBWs or hazard flood warnings (HWs). Results show that IBWs indeed led to higher likelihoods of participants making a protective decision than HWs. Interestingly, we found that key factors identified in previous studies as influential elements in the decision-making process had little impact on their decision. These factors include gender and previous experience with flood warnings. K E Y W O R D Sforecasting and warnings, risk communication, risk perception

The recent extreme rainfall events in Spain such as the storm Gloria have highlighted the gaps in emergency communication, particularly the disconnect between the available impact‐based early warning systems (IBEWSs) and the steps communities take during emergencies. This paper presents a community‐centred framework named ‘site‐specific early warning system’ (SS‐EWS) to co‐design and co‐evaluate with communities an IBEWS for vulnerable locations within high‐risk areas. The components of the framework guide communities in identifying and evaluating local impacts; establishing impact and advisory tables; deriving impact‐based rainfall thresholds and warning levels; and configuring the SS‐EWS with radar‐based nowcasting and numerical weather prediction (NWP) models. A first implementation and evaluation of the SS‐EWS have been done for a public school, two ford crossings and the city of Terrassa, Spain. The SS‐EWS shows promising results in triggering location‐based or site‐specific warnings compatible with the reported impacts and proposing actions to reduce the local risk. Furthermore, the combination of NWP and radar‐based nowcasting improved the capacity of the SS‐EWS to monitor the evolution of the precipitation and capture highly intense rainfall. The SS‐EWS can be a straightforward and cost‐efficient complement for regional EWS to increase the preparedness of communities.

<p><span>The resilience of communities depends on how their citizens react during emergencies and how authorities implement systems to support appropriate self-protection responses from the public. Past events such as the storm Gloria in Spain demonstrate that one of the main challenges in risk communication remains citizens' inappropriate understanding of the upcoming risk and its potential impact in their daily lives. The current official warnings to the population in Spain continue to be based on the exceedances of the event's physical parameters, such as rainfall intensities and accumulations, that can be </span>difficult for citizens to understand, personalize and translate into the expected local risks. The above can create a communication gap<span> between what authorities provide to citizens in terms of risk information and what they actually need from a flood warning to make better decisions and react appropriately during an emergency.</span><strong><span> </span></strong></p><p><span>Society is now demanding localized, people-centred risk communication for better social comprehension and acceptation, which provides understandable information about the expected local impacts and clear guidelines for ensuring citizens' safety during emergencies.</span></p><p><span>Thus, to support citizens' understanding and decision-making process at risk, we present a people-centred approach to design and implement new site-specific warnings (SSWs), i.e. warnings at problematic points based on local vulnerability and exposure information. The proposed methodology places people and communities at the core of the early warning system process. It blends meteorological information coming from radar-based rainfall nowcasting, numerical modelling and historical flood data to translate forecasts into relevant local impacts that the citizens may experience due to the coming weather-induced events and appropriate self-protection actions to help secure their lives. In this context, an active collaboration process with civil protection authorities, stakeholders and citizens is established from the start to incorporate their detailed local knowledge to the system and target their communication needs during emergencies. New technologies, such as smartphone applications, are used to disseminate the SSWs within the area of risk.</span></p><p>A first pilot based on the SSWs methodology is currently at the operational stage in Terrassa, Spain, for selected vulnerable points. Besides contributing to address the current gaps in risk communication, the implemented methodology in this study can help create a proactive, dynamic society by empowering its citizens to respond appropriately during the first instances of an emergency.</p>

<p>Significant progress has been made in the capability and accuracy of forecasting extreme rainfall events and their associated impacts. However, damages remain high and will continue to rise unless immediate actions are taken to support communities in decreasing the impacts of upcoming extreme weather-induced events. In this context, innovative technological tools can help to quickly disseminate relevant impact-based warning information and trigger appropriate self-protection actions based on the local vulnerability and exposure for effective disaster risk reduction.  For this purpose, a mobile app named “A4alerts” has been designed in this research.  </p><p>The tailor-based A4alerts app communicates impact-based warnings for vulnerable locations within high-risk areas (SSWs) generated by a community-based site-specific early warning system (SS-EWS). Based on a participatory approach with community stakeholders, the SS-EWS blends meteorological information coming from radar-based nowcasting, numerical weather prediction models and local risk information to trigger the SSWs disseminated via the A4alerts app. In addition to communicating the active warnings in the area, the app lists the available actions recommended to mitigate and reduce the potential local impacts for each warning level based on pre-approved self-protection plans. Furthermore, users can send geotagged photos and information through the A4alerts app to validate the events and their impacts.</p><p>The A4alerts app has been implemented and tested for selected vulnerable points in cities across Catalonia, Spain. Its capabilities and design have been improved following an iterative approach with end-users to incorporate their feedback and suggestions. Finally, the configuration of the A4alerts app allows it to be easily implemented and exported to new cities to help communities be prepared in times of climate emergency.</p>