Development of a Sensor-Based Flood Early Warning System with SMS Notification in Urban River Restoration Contexts

Urban river restoration areas that function as public spaces and tourism destinations are increasingly exposed to flood risks due to changing hydrological conditions, urbanization, and limited early warning mechanisms. Objective: This study aims to design and implement an automatic flood early warning system based on water level sensors and SMS notifications in the Tukad Tagtag river restoration area, Denpasar, in order to improve flood risk mitigation and public safety. Methodology: This research employs a qualitative applied approach using the Action Research method, which integrates planning, action, observation, and reflection. Data were collected through field observation, system testing, stakeholder consultation, and document analysis. The analysis was conducted using a descriptive and iterative evaluation approach to assess system performance and improvement. Findings: The system successfully detects water level changes in real time, triggers threshold-based alerts, and delivers SMS notifications effectively. System performance improved after iterative adjustments, particularly in sensor positioning and microcontroller configuration, resulting in enhanced stability and reduced false warnings. Implications: The study demonstrates that simple and accessible technology can support effective flood risk mitigation. The system can be applied by local governments and river managers to enhance preparedness, response time, and safety in urban river environments. Originality: This research contributes a practical model by integrating sensor-based detection, real-time processing, and SMS communication within a river restoration tourism context using an iterative Action Research approach.