Floods are prevalent natural calamities that pose a substantial risk to the livelihood, economics, food security and sustainability of a country or region. Over the course of 2023, a total of 160 instances of severe flooding were documented across the globe. Climate change has significantly affected precipitation, a vital element of the hydrological cycle. The resulting fluctuations in precipitation patterns have contributed to the occurrence of severe flood events. India experiences recurrent flood occurrences annually, leading to substantial economic and infrastructural damages. Recent research has observed an increase in the frequency of severe rainfall episodes since 1950. This might be attributed to variations in climate and the rise in human activity. This upward trajectory suggests an increased likelihood of more severe and frequent flood occurrences in the coming future. Approximately 12.5% of India’s territory is susceptible to flooding, with states such as Assam, Bihar, Odisha, Uttar Pradesh and West Bengal being mostly impacted. Forecasting floods have become a formidable task because of the erratic and unpredictable nature of monsoons in India, resulting in periods of excessive rainfall in certain years and abrupt drought-like conditions in others. In order to mitigate the effects, it is imperative for researchers and policymakers to conduct enhanced flood vulnerability assessments to accurately evaluate risks and effectively manage floods. This has led to recent advancements in machine learning (ML), artificial intelligence and artificial neural networks (ANN) combined with geospatial techniques. In this chapter, we showcase the application of multivariate flood characteristics analysis utilising the ‘peak-over-threshold’ approach to evaluate the evolving features and potential process interactions. The approach suggested may be used to evaluate flooding influenced by the impact of natural and human activities in any river basin.