Citizens as Sensors! Integrating the Role of People for Surface Water Flood Mapping by Enhancing Open-Sourced DEM

Increasing frequency and severity of surface water floods are driven by disruption of weather patterns due to climate change, and partly due to land use change from increasing urbanisation. Despite their large societal impact, surface water floods have received less attention compared to other forms of flooding, partly due to the complexity of identifying surface water risks.  Flood mapping and modelling tools used to predict surface water inundation require significant data inputs, which are often unavailable both in terms of resolution and density in resource-limited countries. Though the use of citizen science is witnessed in flood modelling, monitoring, and mapping, these efforts have been mostly limited to validation of the prediction models. Thus, the data gap analysis identified on initial phase of this research highlighted the importance of implementing a citizen science approach to address the gaps in topographic data, which is imperative for flood risk mapping and modelling.

This study adopts a mixed-method approach of qualitative and quantitative analysis to explore the feasibility of citizen-driven data to develop an enhanced Digital Elevation Model (DEM) in a resource-limited, low-income country, Nepal.  DEMs were produced using the geo-coordinates recorded by seventeen community volunteers using their Smartphones under different scenarios using smoothing filters like the Low Pass Filter and Kalman Filter in a GIS interface. The most accurate scenario-based DEM was then utilised to develop a 2D HEC-RAS flood model and flood hazard map for a flood event that occurred in July 2018 in the Hanumante River, Bhaktapur, Nepal. The results were then compared to those produced using the freely available SRTM 30m resolution topographic global dataset.

The study indicates that the accuracy of DEMs created using citizen science and the reliability of the resulting flood risk mapping are shaped by several influences, such as the volunteers’ backgrounds, their motivation levels, the precision of the devices and applications they use to record data, and the safety of the conditions in which data are gathered. Among all participants, students proved to be the most engaged and dependable contributors. The research also showed that directing volunteers to map specific locations leads to higher-quality datasets compared to letting them collect points casually as part of their everyday movements. When collected consistently and with the necessary components, community-driven data can significantly enhance flood risk mapping and modelling. This is especially helpful in data-scarce environments where even minor topographical changes might modify surface water behaviour.

Overall, this study shows that citizen-generated data and community involvement can produce current, affordable topographic data that closes important gaps in conventional datasets. This technique improves local knowledge of terrain characteristics and raises community awareness of surface water flood risk. This demonstrates the wider benefits of citizen science for gathering environmental data, especially in areas where traditional data sources are still scarce.