Iranian Scientists Use Machine Learning Technology for Flood Analysis

“Given the increasing environmental degradation and the need to pay attention to new approaches to assess and monitor the condition of ecosystems, managing natural hazards like floods is necessary,” said Nasiri Khiavi, a PhD graduate in Watershed Science and Engineering from Tarbiat Modarres University.

“Flood sensitivity analysis is one of the most important elements of watershed systems or early warning strategies for preventing and reducing future flood conditions. Because it detects the most vulnerable areas based on physical conditions and determines their tendency to flood. Therefore, the term sensitivity can also be understood as one of the dimensions of flood vulnerability assessment,” he added.

“The results of preparing flood sensitivity and vulnerability maps using a combination of remote sensing techniques, machine learning algorithms, and collaborative game theory can be used by managers, researchers, planners, and even policymakers to identify critical areas and make optimal decisions in watershed science and engineering and water resources management. On the other hand, monitoring most algorithms and optimizing these methods and selecting the best method can save time and apply these methods in future studies,” Nasiri Khiavi said.

Streamflow and flood forecasting remains one of the long-standing challenges in hydrology.

Traditional physically based models are hampered by sparse parameters and complex calibration procedures particularly in ungauged catchments.

4155/v