Climate Meets AI: Unpredictable Changes in Weather Just Got Predictable
The 2021 “AI Earth” Artificial Intelligence Innovation Challenge — AI Helps Accurate Weather and Ocean Forecast is jointly initiated by Nanjing University of Information Science and Technology (NUIST), National Climate Center, National Marine Environmental Forecasting Center (NMEFC), Anhui Meteorology Service, and Alibaba DAMO Academy.
Focusing on the frontier of global atmospheric and oceanic research, the competition aims to promote the application of AI technology to academic cooperation, talent cultivation, technical exchange, and multidisciplinary integration in the field of meteorology and oceanography. By exploring AI technology, the competition aims to break the key bottleneck of the industry, thus developing the prospective research of the industry.
Attracting a total of 2849 teams locally and from overseas, the competition witnessed AI algorithm models developed by 13 teams which surpassed Nature’s 39.2-point paper model standard, indicating more long-term El-Niño forecasts with higher accuracy.
The 2021 “AI Earth” Artificial Intelligence Innovation Challenge went one step further by exploring El-Niño’s 2-year inter-annual predictions while exploring new AI algorithms. AI forecasts show that global climate will stabilize in the next two years, with abnormal weather phenomena like El Niño decreasing in probability. This also means that the probability of flooding in the Yangtze River this upcoming summer is low, ushering in a warmer winter for China towards the end of the year.
The Challenge was championed by a team from Xi’an Jiaotong University led by Weiguang Sang. The winners adopted a brand-new separation spatiotemporal extraction model, alternately using CNN and TCN methods to extract spatiotemporal features, which reduced the number of parameters and the risk of overfitting more effectively.
For the outstanding results achieved in this competition, the Secretary-General of the World Meteorological Organization (WMO) Petteri Taalas also sent congratulations. Professor Taalas stressed how AI is playing an increasingly important role in the field of meteorology and has become the Earth’s “early warning system”. With the basic methods of multi-hazard early warning services, innovation and application of AI are some of the most important areas for WMO and its members in achieving sustainable development goals to better solve future scientific obstacles.
Alibaba DAMO Academy released “AI Earth” — an AI system for multi-source data of sky, space, and ground — at the Apsara Conference 2020. The system is capable of forecasting short-term weather, predicting precipitation within a 1km grid space in just 3 hours. DAMO Academy plans to further make use of AI technology to explore medium-to-long term weather forecasts.