Sand mining activities have accelerated in recent years due to rapid urbanisation, resulting in sand being stripped away from riverbeds of many rivers globally, particularly in Southeast Asia (SEA). While sand mining has caused riverbank erosion, riverbed scouring, and reduced flood frequencies, studies on sand mining rates remain scarce due to difficulties associated with monitoring, either in-situ or long term.
In this project, we seek to enable a step-change in our understanding of riverbed sand mining rates and facilitate sustainable sand harvesting across the rivers in SEA. This project has two overarching aims:
- First, we intend to develop a novel remote sensing-based monitoring system capable of systematically mapping sand mining activities in large rivers. We will use the Mekong River as a “testbed” to implement a field-calibrated Deep Learning model for estimating the sand mining budget of the Mekong River. Our Deep Learning method will subsequently be calibrated with field survey data from ten major rivers across SEA to obtain the first sand mining budget for the entire region.
- Second, the morphodynamic model (MIKE 21C) will be used to identify sediment sinks and investigate the environmental impacts of different sand mining scenarios in all the rivers of SEA. Multiple Criteria Decision-Making will then be used to identify fluvial reaches in SEA where sustainable sand extraction can be carried out.
Our results will provide local governments with the first-ever science-based research to inform regulatory frameworks for sustainable sand mining, as well as provide baseline information for environmental impact assessments. We envision that the novel Deep Learning approach we develop in this project can be adapted to other large rivers across the world, overcoming the methodological challenges associated with monitoring sand mining activities.
Ministry of Education (MOE), Singapore
2023 to 2026