AI flood forecast tool poised to become reality
Source: Chronicle News Service / Premchand Thongam

Imphal, March 27 2025: Artificial intelligence (AI) is gradually becoming an integral part of the modern workforce and daily life.

Amid this global race of AI, a research scholar at the Manipur University's department of computer science is developing an AI tool capable of forecasting floods in the state.

This remarkable initiative is a matter of pride for Manipur, as the project is expected to provide a significant safeguard against the devastating impacts of floods.

Leishangthem Sashikumar of Kakching Wairi Sabal Leikai and a Master of Computer Applications (MCA) degree holder, is pursuing his research at Manipur University.

His research focuses on predicting potential rise in water levels that could lead to floods by analysing various parameters such as rainfall patterns, weather conditions, relative humidity, and fluctuations in water levels in the major river systems of the state.

Sashikumar's model utilises machine learning and deep learning techniques, which are components of artificial intelligence, to process and analyse these parameters.

Speaking to The People's Chronicle, Sashikumar explained that his research is based on a case study of Manipur, where he uses daily water level data from the state's four major rivers-Imphal, Nambul, Iril, and Thoubal.

These data points are collected from monitoring stations at Thong Nambonbi, Minuthong, Lilong Thong, Irilbung, and Ithai Barrage under the state's water resources department.

Additionally, he incorporates rainfall data obtained from the Directorate of Environment as part of his research.

Sashikumar has meticulously gathered data on rainfall activities, water levels, daily weather, and relative humidity of the past six years.

These datasets serve as key parameters for training his AI model through machine learning and deep learning applications.

His model can predict the likelihood of a flood in the next three days by analysing relevant data from the preceding five days.

The AI tool generates a potential water level based on these inputs, and the algorithm subsequently issues an early flood warning.

The threshold values used for triggering the early warning system are based on water levels set in flood-prone areas, as provided by the Water Resources Department.

During the testing phase, Sashikumar's AI model achieved an impressive efficiency level of 0.97.In AI applications of this nature, an efficiency level of 1 is considered highly accurate, making his model's performance almost equivalent to achieving 100 per cent accuracy.

Although the model falls just short of the perfect efficiency mark, Sashikumar emphasised that it can still be considered highly reliable.

While the model is nearly complete, he estimates that it could take approximately one year to develop a fully functional AI application in the form of a website.

If successful, the tool could become a vital asset for the state in the years to come, he said.

The need for such an AI tool is critical, as the water levels of the state's major rivers often rise to flood levels during the rainy season, leading to breaches of riverbanks and subsequent floods.

These floods have devastating consequences, impacting the lives and properties of people residing in flood-prone areas.

In recent years, flooding has become a recurring issue, posing significant challenges for disaster management.

Having an AI-powered flood forecasting tool could prove to be invaluable in mitigating these risks by enabling people in vulnerable areas to take timely preventive measures and safeguard their lives and properties.

Sashikumar's initiative not only showcases the potential of AI in addressing real-world challenges but also highlights the contribution of a local scholar in developing technology that could have a profound impact on disaster management in the state.

The successful implementation of his AI model could pave the way for more innovative solutions that enhance the resilience of communities in the face of natural disasters.

(This article is published under the 5th Scientific Journalism Program on 'Artificial Intelligence and Its Usage' of the Public Relations and Media Management (PRMM) Cell, Central Agricultural University, Imphal) .

Comment