SCIENCE@home

Machine learning modelling of flood susceptibility in a tropical river basin of South India

N. Anjali,Girish Gopinath

Published 2026 in International Journal of Disaster Studies and Climate Resilience

ABSTRACT

Flooding is a major recurring hazard in tropical monsoon climates, causing extensive damage to lives, infrastructure, and economic activities. The Achankovil River Basin in southern India is highly susceptible to recurrent monsoon-induced flooding, particularly in its low-lying downstream reaches. This study assesses flood susceptibility in the Achankovil River Basin using a machine-learning Maximum Entropy (MaxEnt) model and identifies the key environmental factors controlling flood occurrence. Ten geo-environmental variables, such as geomorphology, soil texture, Vertical Distance to Channel Network (VDCN), slope angle, rainfall, land use/land cover (LU/LC), distance from streams, Topographic Wetness Index (TWI), upslope curvature, and downslope curvature, were used together with a detailed flood inventory comprising 677 flood affected locations. The model is validated using the Receiver Operating Characteristic (ROC) curve and area under the curve (AUC) value (ROC–AUC). The predicted flood susceptibility map is further grouped into low, moderate, high, and very high susceptibility regions. The results show that high and very high flood susceptibility zones are mainly concentrated in the downstream and central lowland regions along major river channels, characterised by low elevation, gentle slopes, and proximity to drainage networks, which favours inundation. In contrast, the eastern highland sector of the study area exhibits predominantly low flood susceptibility zones. The model validation returns an AUC value of 0.89, indicating good model accuracy and reliability. Besides, the model identified Vertical Distance to the Channel Network (VDCN) as the most important factor, followed by rainfall and slope angle. The findings confirm the robustness of the maximum entropy model and provide valuable insights for flood early warning and land-use development in the study area.

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-42 of 42 references · Page 1 of 1

CITED BY

  • No citing papers are available for this paper.

Showing 0-0 of 0 citing papers · Page 1 of 1