August 12th 2024.
The Indian Institute of Technology in Bhubaneswar has made a groundbreaking achievement in the field of weather prediction. According to a statement released on Monday, the institute has successfully developed a technology that can accurately forecast rainfall, particularly in cases of heavy downpours, with enough warning time.
This breakthrough was made possible by combining the output of the Weather Research and Forecasting model with a deep learning model. The institute has conducted extensive studies using past cases in the challenging terrain of Assam during June 2023 and in the state of Odisha, where heavy rainfall events are unpredictable due to the arrival of multiple intense rain-bearing monsoon low-pressure systems.
One such case was in June 2023, when Assam was hit by severe flooding due to heavy rainfall. The DL model proved to be significantly more accurate in predicting the intensity and distribution of rainfall at a district level. The institute proudly proclaims that the model was able to accurately predict the spatial distribution and intensity of rainfall, showcasing its remarkable performance.
The research team utilized the WRF model to generate real-time initial weather forecasts, which were then further refined using the DL model. This approach allowed for a more comprehensive analysis of rainfall patterns, incorporating a special attention module to capture the intricate spatial dependencies in the data.
The results were astounding, with the model displaying nearly double the prediction accuracy of traditional ensemble models at a district level with a lead time of up to 96 hours. It is a significant advancement in the field of weather forecasting, and the institute hopes that it will help in mitigating the impact of severe weather events on vulnerable communities.
PTI
[This article has been trending online recently and has been generated with AI. Your feed is customized.]
[Generative AI is experimental.]
