Sydney (ANTARA) \u2014 An international team of researchers has developed\nan artificial intelligence (AI) model to support long-range flood\nforecasting, including by helping communities receive disaster warnings\nseveral weeks before an event.<\/p>\n
The researchers used two decades of historical precipitation and\nriver-flow data from the DuPage River watershed in Illinois, the United\nStates, to develop a deep-learning model incorporating an attention\nmechanism to identify rainfall and river-flow patterns, according to a\nstatement released by the University of Technology Sydney (UTS) in\nAustralia on Tuesday (13 October).<\/p>\n
By combining predictions from several models, the team improved\nforecast accuracy. The ensemble deep-learning model reduced prediction\nerrors by about 38 per cent, while the proposed fusion framework\nimproved predictive performance by around 47 per cent, according to the\nstudy published in the Journal of Hydrologic Engineering.<\/p>\n
The approach remained highly accurate for forecasts up to 12 weeks\nahead, said Professor Amir H. Gandomi of UTS, one of the study\u2019s\nauthors, emphasising that the additional warning time could help\ncommunities prepare for floods.<\/p>\n
The findings could have implications for flood-risk management,\nincluding disaster preparedness, infrastructure resilience and\nwater-resource management, the research team said.<\/p>\n
\u201cUltimately, the goal is not simply to predict floods more\naccurately, but to turn better predictions into better decisions,\u201d\nGandomi said.<\/p>", "url": "https:\/\/jawawa.id\/newsitem\/researchers-develop-ai-model-to-support-long-range-flood-forecasting-1791896057", "image": "" }, "sponsor": "Okusi Associates", "sponsor_url": "https:\/\/okusiassociates.com" }