Bhubaneswar: The Indian Institute of Technology (IIT), Bhubaneswar has developed a hybrid technology integrating the output from the Weather Research and Forecasting (WRF) and Deep Learning (DL) models to enhance the prediction accuracy of heavy rainfall with an adequate lead time.
This model will lend more credence to weather forecasting at a time when heavy rainfalls triggered by low-pressure system and climate change with devastating outcomes are increasingly becoming common across India, said the group of researchers of IIT Bhubaneswar which developed the technology.
The team comprising Dr. Dhananjay Trivedi, Dr. Omveer Sharma, Dr. Vivekananda Hazra, Dr. Sandeep Pattnaik from the School of Earth Ocean and Climate Sciences and Dr. Niladri Bihari Puhan of School of Electrical and Computer Sciences at IIT Bhubaneswar studied the complex terrain of Assam which is highly vulnerable to severe flooding and Odisha where heavy rainfalls are highly dynamic due to the landfall of multiple and intense rain-bearing monsoon low-pressure systems.
The work has been supported by the Council of Scientific and Industrial Research (CSIR), and the New Venture Fund U.S., and IIT Bhubaneswar.
The study titled ‘Minimization of Forecast Error Using Deep Learning for Real-Time Heavy Rainfall Events Over Assam’ has been published in the IEEE Xplore, a research database.
Between June 13 and 17, 2023, Assam witnessed severe floods due to heavy rainfall. The DL model was able to more accurately predict the distribution and intensity of rain. The research employed the WRF model to generate initial weather forecasts real time, which were then refined using the DL model, said the IIT Bhubaneswar team.
This technology allowed more detailed analysis of rainfall patterns. To improve accuracy, the model was trained using data from past heavy rainfall events and observations of India Meteorological Department (IMD).
In Assam, the hybrid model displayed doubled prediction accuracy of that of traditional ensemble models at a district level with a lead time up to 96 hours.
The researchers from the IIT Bhubaneswar demonstrated a significant leap in accurately predicting heavy rainfall events in the complex terrain of Assam in real-time situations.
The study has revealed that integrating DL technology with the traditional WRF model dramatically improves forecast accuracy for heavy rainfall events in real-time, a critical advancement for the flood-prone mountainous region such as Assam.
Key findings of the study
District-Level Precision: First of its kind in real time to improving forecast skills on a district-scale.
Enhanced Prediction Accuracy: The DL model demonstrated a notable improvement in forecast accuracy, capturing 54.4% of heavy rainfall events compared to the WRF model’s 22.8%. The DL model achieved a mean absolute error (MAE) of under 30 mm, significantly lower than WRF’s more than 50 mm MAE.
Technological Innovation: The research introduces a U-Net model with a spatio-attention (SA) module that captures intricate spatial dependencies of rainfall features at district scale.
These findings demonstrate the immense potential of Artificial Intelligence (AI) in improving real-time weather forecasting, particularly for heavy rainfall events in the complex terrains in India. This advancement is crucial for mitigating the impacts of natural disasters and ensuring public safety. These works will help create analogous hybrid models for other intricate topographical terrain areas such as the Western Himalayas and Western Ghats regions of India, the research team of IIT Bhubaneswar said.